platform/upstream/efl.git
4 years agoelput - fix #else path for minor
Carsten Haitzler (Rasterman) [Wed, 27 May 2020 15:24:46 +0000 (16:24 +0100)]
elput - fix #else path for minor

4 years agoelput - it seems not everyone has major and minor macros defined
Carsten Haitzler (Rasterman) [Wed, 27 May 2020 14:30:59 +0000 (15:30 +0100)]
elput - it seems not everyone has major and minor macros defined

so handle those cases and provide some implementation for those,
otherwise use the macros.

4 years agoecore_wl2_window: copy available rotation info.
Hosang Kim [Wed, 27 May 2020 14:05:30 +0000 (10:05 -0400)]
ecore_wl2_window: copy available rotation info.

Summary:
ecore_wl2_window doesn't copy rotation information.
If passed pointer is destroyed, rotation information disappears.
So add memory allocating for managing information.

Reviewers: CHAN, devilhorns

Reviewed By: devilhorns

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11891

4 years agoecore_wl2: Fix event type name.
Woochanlee [Wed, 27 May 2020 14:03:38 +0000 (10:03 -0400)]
ecore_wl2: Fix event type name.

Summary: It's been missed commit 9ca13ef5c10904b620407ce8981fadf2a903b180

Reviewers: Hermet, devilhorns

Reviewed By: devilhorns

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11889

4 years agobuild: efl-one track all subdirs correctly
Marcel Hollerbach [Wed, 27 May 2020 12:38:32 +0000 (14:38 +0200)]
build: efl-one track all subdirs correctly

this was missed before, so the subdirectories have not been correctly
added. With this you can now simply build with efl-one with including
<Elementary.h>

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11893

4 years agoedje: fix memory leak in edje part recalculation.
Myoungwoon Roy, Kim [Wed, 27 May 2020 00:03:15 +0000 (00:03 +0000)]
edje: fix memory leak in edje part recalculation.

Allocated part_array is only freed when _circular_dependency_find() is successful. It is not freed in other cases.

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Reviewed-by: woochan lee <wc0917.lee@samsung.com>
Differential Revision: https://phab.enlightenment.org/D11886

4 years agoEina file test: use the correct Eina file object
Vincent Torri [Tue, 26 May 2020 21:48:56 +0000 (21:48 +0000)]
Eina file test: use the correct Eina file object

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D11885

4 years agoEina file test: on Windows, open cmd, not a shell
Vincent Torri [Tue, 26 May 2020 21:45:31 +0000 (21:45 +0000)]
Eina file test: on Windows, open cmd, not a shell

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D11884

4 years agobuild - systemd service files - install fix if systemd not found
Carsten Haitzler (Rasterman) [Wed, 27 May 2020 13:10:33 +0000 (14:10 +0100)]
build - systemd service files - install fix if systemd not found

so if systemd is not found - installe in PREFIX/lib/systemd/user which
allows build to work without disabling systemd.

4 years agobuild: add correct flags to efl-one.pc
Marcel Hollerbach [Wed, 27 May 2020 11:09:36 +0000 (13:09 +0200)]
build: add correct flags to efl-one.pc

they are required, and normally dragged in via eina.

4 years agobuild - fix deps for efl_canvas_wl to be able to build against efl
Carsten Haitzler (Rasterman) [Wed, 27 May 2020 10:45:41 +0000 (11:45 +0100)]
build - fix deps for efl_canvas_wl to be able to build against efl

4 years agobuild: stop buildsystem from beeing a public dependency
Marcel Hollerbach [Tue, 26 May 2020 08:53:45 +0000 (10:53 +0200)]
build: stop buildsystem from beeing a public dependency

this is wrong, each library should declare it on its own

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11883

Change-Id: Idb9f561caa2707536e42faacc651005ed4587a22

4 years agobuild: time for efl-one
Marcel Hollerbach [Sun, 10 May 2020 12:43:28 +0000 (14:43 +0200)]
build: time for efl-one

this is building parts of efl into a single .so and links all modules
and binaries to it.

The libraries themselfs are build as .a's which are linked together as
.so's. Which is required as every subproject has its little custom
c_flags, which are somtimes conflicting.

After the final .so is then built, all the split up libraries are
replaced with the efl-one. After that the modules and binaries are built
correctly with the correct link on the efl-one parts.

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11808

Change-Id: Icbb8b0f60c2581e07c57e87ae66f44bc8c0e7de8

4 years agorefactor build
Marcel Hollerbach [Sun, 10 May 2020 12:41:57 +0000 (14:41 +0200)]
refactor build

libraries are split into deps, external deps, and pub deps.
Evas engines are refactored to use the predefined engine deps.

this is preparation work for efl-one.

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11806

Change-Id: Ib08b551dd830ac4d69dda029442ea0bced2af30e

4 years agobuild: split up building of libs, bins, modules, benchmark
Marcel Hollerbach [Thu, 7 May 2020 17:05:54 +0000 (19:05 +0200)]
build: split up building of libs, bins, modules, benchmark

this is needed as a preparation for more build refactoring. Overall goal
to build efl in a single .so.

What this commit results in is that *every* lib gets build first, then
the modules and binaries of all subprojects. This way of building is
needed in order to replace the split up libraries with the efl-one one
later.

Additionally, the infrastructure is added to set flags on the libraries
that should be build together.

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11805

Change-Id: Ie85c840aa6280687500223d97fdeb001d432114e

4 years agoefl gfx_path: fix memory leak.
Hermet Park [Wed, 27 May 2020 06:34:24 +0000 (15:34 +0900)]
efl gfx_path: fix memory leak.

his reverts commit 38be95b0b6fd93df01636ae0a7c974b32ed694a1.

The reverting patch blocks efl_canvas_vg_shape triggers efl_canvas_vg_node destructor.

That occurs massive leaks among the all nodes in the vector tree.

To fix the origin issue, we must reset path explicitly in the derived class.

@fix

4 years agoedje: fix memory leak by freeing the limits pointer.
Subhransu Mohanty [Wed, 27 May 2020 05:55:44 +0000 (14:55 +0900)]
edje: fix memory leak by freeing the limits pointer.

Summary:
This limit parts are only used in naviframe edc thats why it is
only showing up in naviframe test app.

Reviewers: Jaehyun_Cho, Hermet

Reviewed By: Jaehyun_Cho

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11887

4 years agoedje_load : clean up vector resource when edje file freed.
Woochanlee [Wed, 27 May 2020 03:32:11 +0000 (12:32 +0900)]
edje_load : clean up vector resource when edje file freed.

Summary: Fix memory leak.

Reviewers: cedric, Hermet, smohanty, bu5hm4n

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11859

4 years agoedje: Fix memory leak by freeing the filter data.
Subhransu Mohanty [Wed, 27 May 2020 03:31:35 +0000 (12:31 +0900)]
edje: Fix memory leak by freeing the filter data.

Reviewers: Hermet, CHAN

Reviewed By: CHAN

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11879

4 years agoecore-evas-wayland: fix migration build error
Jongmin Lee [Tue, 26 May 2020 22:29:01 +0000 (07:29 +0900)]
ecore-evas-wayland: fix migration build error

Change-Id: Ieb0ac38990ddebad4c9b539712fb7e04e637b066

4 years agoecore-evas-wayland: Fix unused variables
Christopher Michael [Tue, 26 May 2020 18:51:53 +0000 (14:51 -0400)]
ecore-evas-wayland: Fix unused variables

Patch 09ab8c9967512e67b7 added code to printf wl client animation
ticks, however the line that actually does the printing is commented
out (which is fine) but the variables used in that printf are not
needed if we are not printing things out

4 years agoecore-wl2: Remove unused variables
Christopher Michael [Tue, 26 May 2020 18:49:32 +0000 (14:49 -0400)]
ecore-wl2: Remove unused variables

Patch d584696f12c6d93e26 remove the unmapping of subsurfaces on window
hide, however the variables were left declared so remove those

4 years agoecore-wl2: Fix build break from patch 9ca13ef5c10904b620407
Christopher Michael [Tue, 26 May 2020 18:47:56 +0000 (14:47 -0400)]
ecore-wl2: Fix build break from patch 9ca13ef5c10904b620407

Above commit caused a build break from a missing comma ... Sole blame
falls on the reviewer for not doing a build test ... bad me ;)

4 years agoecore_wl2 : Organize window creation/deletion events.
Woochanlee [Tue, 26 May 2020 15:49:16 +0000 (11:49 -0400)]
ecore_wl2 : Organize window creation/deletion events.

Summary:
Add create / destroy events.
Call hide event when the window terminated if the window was visible.

Reviewers: devilhorns, Hermet, raster

Reviewed By: devilhorns

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11878

4 years agoecore_wl2_window: do not unmap subsurface in hide
Shinwoo Kim [Tue, 26 May 2020 15:49:07 +0000 (11:49 -0400)]
ecore_wl2_window: do not unmap subsurface in hide

Summary:
If window_hide should unmap subsurface, then window_show should re-map
the unmapeed subsurface. I have no idea why window_hide unmap all subsurfs.

Reviewers: zmike, devilhorns

Reviewed By: zmike, devilhorns

Subscribers: devilhorns, cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11869

4 years agoevas_textblock: fix doc typo
Ali Alzyod [Mon, 18 May 2020 07:47:58 +0000 (07:47 +0000)]
evas_textblock: fix doc typo

Reviewed-by: Xavi Artigas <xavierartigas@yahoo.es>
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11845

4 years agobuild: disable leak detection in address sanitizer build for tests as well
Stefan Schmidt [Tue, 26 May 2020 13:54:57 +0000 (09:54 -0400)]
build: disable leak detection in address sanitizer build for tests as well

Summary:
The same principle we use for the build stage we use here for running
our tests. Leak detection has just to many problems for us on the
shutdown path that it makes no sense to use here.

I am also bumping the evas timeout form the default 30s to 60 as I have
been running into timeouts with asan enabled on my machine. This would
not change anything on a default build.
Depends on D11137

Reviewers: smohanty, bu5hm4n, raster, zmike

Reviewed By: zmike

Subscribers: zmike, cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11862

4 years agoecore_main_timechanges: fix include
Marcel Hollerbach [Mon, 25 May 2020 09:24:28 +0000 (11:24 +0200)]
ecore_main_timechanges: fix include

we should not prefix this with the full path.

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Reviewed-by: Vincent Torri <vincent.torri@gmail.com>
Differential Revision: https://phab.enlightenment.org/D11881

4 years agoevas - csd and wl csd - fix alpha zeroing to avoid framespace junk
Carsten Haitzler (Rasterman) [Tue, 26 May 2020 11:16:09 +0000 (12:16 +0100)]
evas - csd and wl csd - fix alpha zeroing to avoid framespace junk

i noticed some junk in framespace... this happesn when solid objects
are outside the normal window content area and their solid areas are
not clipped to this inner frame area thus those areas are not "zeroed
out" next render.

@fix

4 years agocanvas vg: fix vector file data memory leak.
Hermet Park [Tue, 26 May 2020 09:47:44 +0000 (18:47 +0900)]
canvas vg: fix vector file data memory leak.

this file data must be freed when eet file is closed.

@fix

4 years agoevas vg: clean up dangling instances
Hermet Park [Tue, 26 May 2020 09:37:41 +0000 (18:37 +0900)]
evas vg: clean up dangling instances

unref containers when file is closed.

@fix

4 years agoMerge efreet_mime efreet_trash into efreet
Marcel Hollerbach [Sun, 10 May 2020 10:47:58 +0000 (12:47 +0200)]
Merge efreet_mime efreet_trash into efreet

efreet_mime and efreet_trash are now included in libefreet.so. There are
compatible libs for libefreet_mime.so libefreet_trash.so, which simple
drag in libefreet.so. .pc files also are in place for compatibility.

efreet_init & shutdown now initializes and shutdowns trash and mime.
The old init functions now simply call efreet_init, efreet_init will
then init the trash or mime libs.

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11807

4 years agolib: add more c_args handling for places we missed before
Stefan Schmidt [Tue, 19 May 2020 14:30:08 +0000 (16:30 +0200)]
lib: add more c_args handling for places we missed before

These meson files did not have the c_args handling before. Make sure we
use package_c_args here as well.

Reviewed-by: Vincent Torri <vincent.torri@gmail.com>
Reviewed-by: João Paulo Taylor Ienczak Zanette <joao.tiz@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11860

4 years agobuild: modules: harmonize use of package_c_args
Stefan Schmidt [Mon, 18 May 2020 15:58:28 +0000 (17:58 +0200)]
build: modules: harmonize use of package_c_args

Adding it to all missing places so we can control it from the central
meson.build file.

Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Reviewed-by: João Paulo Taylor Ienczak Zanette <joao.tiz@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11855

4 years agobuild: lib: harmonize the use of package_c_args in all libs
Stefan Schmidt [Mon, 18 May 2020 15:46:55 +0000 (17:46 +0200)]
build: lib: harmonize the use of package_c_args in all libs

Add it to subprojects which are not using it and remove and old
ELEMENTARY_BUILD define we no longer use. This allows us to have a
central place in the main meson.build file to set this variable.

Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Reviewed-by: Vincent Torri <vincent.torri@gmail.com>
Reviewed-by: João Paulo Taylor Ienczak Zanette <joao.tiz@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11854

4 years agobenchmarks: use EFL_BUILD=1 in c_args
Stefan Schmidt [Mon, 18 May 2020 15:24:13 +0000 (17:24 +0200)]
benchmarks: use EFL_BUILD=1 in c_args

Original patch by Vincent Torri.

Reviewed-by: Vincent Torri <vincent.torri@gmail.com>
Reviewed-by: João Paulo Taylor Ienczak Zanette <joao.tiz@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11853

4 years agobuild: move EFL_BUILD to package_c_args used in all subprojects
Stefan Schmidt [Mon, 18 May 2020 15:18:54 +0000 (17:18 +0200)]
build: move EFL_BUILD to package_c_args used in all subprojects

Original patch by Vincent Torri. Co-authored with Marcel Hollerbach.

fixup

Reviewed-by: Vincent Torri <vincent.torri@gmail.com>
Reviewed-by: João Paulo Taylor Ienczak Zanette <joao.tiz@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11851

4 years agocanvas vg: fix dangling vector containers.
Hermet Park [Tue, 26 May 2020 07:08:30 +0000 (16:08 +0900)]
canvas vg: fix dangling vector containers.

These duplicated containers are not going removed
since its refernece is always more than 0.

We need to unref explicity when they have to

@fix

4 years agoedje: fix animation memory leak.
Hermet Park [Tue, 26 May 2020 05:37:10 +0000 (14:37 +0900)]
edje: fix animation memory leak.

if object is destructed on canvas animation,
its internal resource is leaked.

Here simple fix brings stopping animation before deleting obj.

@fix

4 years agoRevert "evas image: Support android 9 patch feature."
Hermet Park [Tue, 19 May 2020 03:34:05 +0000 (12:34 +0900)]
Revert "evas image: Support android 9 patch feature."

This reverts commit 23d4e94b3479f7e557c13f9557f43dd765c6d7b5.

9 patch feature was introduced by an another approach.

Change-Id: I8ce722d6a1163302d41dfc3215dd7a8850f997d2

4 years agolibs: headers: ensure we include headers in the correct order for EAPI
Stefan Schmidt [Mon, 18 May 2020 16:07:31 +0000 (18:07 +0200)]
libs: headers: ensure we include headers in the correct order for EAPI

We need to ensure this order to make sure EAPI is working correctly on
windows.

Original patch by Vincent Torri.

Differential Revision: https://phab.enlightenment.org/D11856

Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
4 years agobenchmark: fix ident in meson.build file
Stefan Schmidt [Mon, 18 May 2020 15:22:12 +0000 (17:22 +0200)]
benchmark: fix ident in meson.build file

Use two spaces indent as in other files.

Differential Revision: https://phab.enlightenment.org/D11852

Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
4 years agobuild: set DLL_EXPORT when on a windows system
Stefan Schmidt [Mon, 18 May 2020 14:53:22 +0000 (16:53 +0200)]
build: set DLL_EXPORT when on a windows system

No need to have this elementary specific, we can do this globally.

Original patch by Vincent Torri.

Differential Revision: https://phab.enlightenment.org/D11850

Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
4 years agoeina - dont use SCHED_BATCH or SCHED_IDLE unless they are defined
Carsten Haitzler (Rasterman) [Mon, 25 May 2020 09:27:19 +0000 (10:27 +0100)]
eina - dont use SCHED_BATCH or SCHED_IDLE unless they are defined

4 years agoecore_evas_wayland_common: keep the backword compatibility for wayland dnd handler
Taehyub Kim [Mon, 18 May 2020 07:56:08 +0000 (07:56 +0000)]
ecore_evas_wayland_common: keep the backword compatibility for wayland dnd handler

Keep the backword compatibility for wayland ECORE_WL2_EVENT_DATA_SOURCE_END event handler.
If we don't use this handler, the dnd is not ended when the drop item is dropped.
@fix

Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11846

4 years agoelm_dnd: do not register widgets twice
Marcel Hollerbach [Mon, 18 May 2020 17:04:43 +0000 (19:04 +0200)]
elm_dnd: do not register widgets twice

we have to support none widgets, but we should not register widgets
twice because of that.

Differential Revision: https://phab.enlightenment.org/D11877

4 years agoevas image: reload after alpha_set
Shinwoo Kim [Mon, 25 May 2020 05:48:31 +0000 (14:48 +0900)]
evas image: reload after alpha_set

Summary:
The alpha_set cancels preload, and do not try to load it again.
So the image is not showing, if the alpha_set is called while preloading.

[Sample]
   Evas_Object *image = elm_image_add(box);
   elm_box_pack_end(box, image);
   evas_object_show(image);

   elm_image_file_set(image, "img.png", NULL);     // start preloading

   Evas_Object *eimg = elm_image_object_get(image);
   evas_object_image_alpha_set(eimg, EINA_TRUE);   // cancel preloading

Reviewers: Hermet, jsuya, herb

Reviewed By: Hermet

Subscribers: zmike, cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11864

4 years agoedje: markup flag should be TRUE in legacy edje_object_part_text_set
Bowon Ryu [Fri, 22 May 2020 06:27:25 +0000 (15:27 +0900)]
edje: markup flag should be TRUE in legacy edje_object_part_text_set

Summary:
Through some APIs (elm_entry_editable_set, elm_entry_single_line_set, etc.)
located after edje_object_part_text_set(legacy)

If efl_ui_widget_theme_apply is called,
In edje_object_part_text_raw_generic_set, the legacy flag becomes FALSE.
And in this case, the logic works in the unintended direction
because the set_markup flag is FALSE.

Test Plan:
/*
gcc -o entry_example entry.c `pkg-config --cflags --libs elementary`
 */

EAPI_MAIN int
elm_main(int argc EINA_UNUSED, char **argv EINA_UNUSED)
{
   Evas_Object *win, *en;

   elm_policy_set(ELM_POLICY_QUIT, ELM_POLICY_QUIT_LAST_WINDOW_CLOSED);

   win = elm_win_util_standard_add("entry-example", "test");
   elm_win_autodel_set(win, EINA_TRUE);
   en = elm_entry_add(win);
   evas_object_size_hint_weight_set(en, EVAS_HINT_EXPAND, EVAS_HINT_EXPAND);
   elm_object_part_text_set(en, "elm.guide", "<font_size=32 color=#FFF>TEST</font_size>");
   elm_entry_editable_set(en, EINA_FALSE);

   evas_object_show(en);

   elm_object_content_set(win, en);
   evas_object_resize(win, 300, 200);
   evas_object_show(win);

   elm_run();

   return 0;
}
ELM_MAIN()

Reviewers: woohyun, ali.alzyod

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11868

Change-Id: Ic65a0ca0df749f6f4feb0a21b7bccb62c70345f4

4 years agoecore_evas: Fix doc.
Woochanlee [Fri, 22 May 2020 05:06:52 +0000 (14:06 +0900)]
ecore_evas: Fix doc.

Deprecated tag should be on top of the API doc.

Change-Id: I662f2ee67b6a3dd192399d64a91de50a86a89c1a

4 years agoewk-wrapper: change base class
Bartlomiej Grzelewski [Fri, 15 May 2020 09:29:44 +0000 (11:29 +0200)]
ewk-wrapper: change base class

EWK_WRAPPER_CLASS inherits after EFL_UI_WIDGET_CLASS now.

This patch allows ewk content to be visible in atspi tree.

Change-Id: Ie90c20dbd2fc3dc7b4f2f641d5d77377f670b9da

4 years agoatspi: parsing fix
Bartlomiej Grzelewski [Fri, 8 May 2020 12:11:00 +0000 (14:11 +0200)]
atspi: parsing fix

Old implementation parsed string: ":1.23:/object_path" into:
".1.23" and "/object_path"
Expected:
":1.23" and "/object_path"

This commit reverts change made by commit:
7e26dcf224658f590e0b8460548c37f659d828f9

Change-Id: I69f60dbda6df1fb1f755540543000d7340e39a37

4 years agoecore_evas : Add deprecated Tags and macro for deprecated APIs.
Woochanlee [Thu, 21 May 2020 04:51:08 +0000 (13:51 +0900)]
ecore_evas : Add deprecated Tags and macro for deprecated APIs.

Deprecated below APIs.

ecore_evas_comp_sync_get/set
ecore_evas_app_comp_sync_get/set

Change-Id: Iebd2cd4cdfb571433b0f7da155b8ab1d5e2bda9a

4 years agoopengl: Fix build break for desktop gl backend
Subhransu Mohanty [Thu, 21 May 2020 04:48:20 +0000 (13:48 +0900)]
opengl: Fix build break for desktop gl backend

The desktop gl backend was broken because of code addition related to TBM and egl .
fix the issue by putting them under GL_GLES flag.

Change-Id: Idb1ad0ee06bfa45e6cd31ba2f00654d3a3fa9434

4 years agoelm_entry: do not execute entry theme apply before finalized
Bowon Ryu [Wed, 20 May 2020 05:59:54 +0000 (14:59 +0900)]
elm_entry: do not execute entry theme apply before finalized

theme apply is called in layout's finalize function.
therefore, entry theme apply is called during finalize.
for performance, it is fixed not to execute entry theme apply before
finalized.

@tizen_only

Change-Id: I222de8020fb90cdcbefce9eadbe2e93c43b2b074
Signed-off-by: Bowon Ryu <bowon.ryu@samsung.com>
4 years agonaviframe: Fix buildbreak for X11 backend by putting wayland specific code under...
Subhransu Mohanty [Wed, 22 Apr 2020 06:54:52 +0000 (15:54 +0900)]
naviframe: Fix buildbreak for X11 backend by putting wayland specific code under Wayland guard.

Change-Id: Iec8a8bf8e25909c128ddd7f800b39d5a95cf3bdb

4 years agoelm test: include evil_private.h to declare realpath() on Windows
Vincent Torri [Sat, 23 May 2020 13:57:10 +0000 (14:57 +0100)]
elm test: include evil_private.h to declare realpath() on Windows

Reviewers: raster, zmike

Reviewed By: raster

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11872

4 years agoecore_evas: fix uninitialized variable warning
Vincent Torri [Sat, 23 May 2020 13:55:53 +0000 (14:55 +0100)]
ecore_evas: fix uninitialized variable warning

Test Plan: compilation

Reviewers: raster, zmike

Reviewed By: raster

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11871

4 years agoecore - another systemd header - go away. don't need it now.
Carsten Haitzler (Rasterman) [Sat, 23 May 2020 08:41:38 +0000 (09:41 +0100)]
ecore - another systemd header - go away. don't need it now.

4 years agoecore evas - tick printf debugs - disabled right now so no noise
Carsten Haitzler (Rasterman) [Sat, 23 May 2020 08:33:54 +0000 (09:33 +0100)]
ecore evas - tick printf debugs - disabled right now so no noise

this is here so it can be turned on and off to see what's happening
with ticking and tick latency

4 years agoecore-x : move x vsync device back to mainloop from thread
Carsten Haitzler (Rasterman) [Sat, 23 May 2020 08:28:45 +0000 (09:28 +0100)]
ecore-x : move x vsync device back to mainloop from thread

after i noticed some jitters, it seesms one thread to listen for vsync
events then to wake up the mainloop can suffer from irregular
scheduling jitters. it seems to be highly depenedent on both gpu
driver and cpu but it seemed the vsync thread itself was more reliably
woken than the mainlooop it then signalled, so merging this back is
just batter. it's configurable via an environment variable so we can
try either right now and see, but default is now to be in main loop.

4 years agoecore thread - feedback threads should not be background threads...
Carsten Haitzler (Rasterman) [Thu, 21 May 2020 17:00:47 +0000 (18:00 +0100)]
ecore thread - feedback threads should not be background threads...

if try_no_queue is used - its a thread that clearly wantt to be out of
the queue to run on its own so probably wants to wake up accurately
and thus not be a backgroun rpriority task but and urgent one. ensure
we set up priority accordingly as we didn't before. this should fix
scheduling when under load for vsync

@fix

4 years agoecore-x expose some debug times like the exact time the vsync woke
Carsten Haitzler (Rasterman) [Thu, 21 May 2020 16:59:53 +0000 (17:59 +0100)]
ecore-x expose some debug times like the exact time the vsync woke

this is for finding and eliminating jank in the vsync scheduling etc.
- as its underscore it's clearly a private api.

4 years agoecore - expose "hidden" internal api for debugging to get wakeup time
Carsten Haitzler (Rasterman) [Thu, 21 May 2020 16:58:46 +0000 (17:58 +0100)]
ecore - expose "hidden" internal api for debugging to get wakeup time

get the real time at which main loop woke up for purposes of debugging
and finding jank/scheduling jitter.

4 years agoevas_vg_container: return NULL in evas_vg_container_add when the parent of container...
Taehyub Kim [Fri, 22 May 2020 03:23:46 +0000 (12:23 +0900)]
evas_vg_container: return NULL in evas_vg_container_add when the parent of container is NULL

Summary:
evas_vg_container should return NULL when the parent is NULL
because the function does not allot NULL parent
@fix

Reviewers: kimcinoo, jsuya, Hermet

Reviewed By: kimcinoo

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11867

4 years agoelua: prepare for compatibility with cffi-lua (use ffi.nullptr)
Daniel Kolesa [Thu, 21 May 2020 15:45:43 +0000 (17:45 +0200)]
elua: prepare for compatibility with cffi-lua (use ffi.nullptr)

This is necessary because the LuaJIT FFI implements the null
pointer to nil equality behavior, which is not possible to
do in standard Lua (as equality metamethod needs identical types
to trigger).

So replace it with ffi.nullptr as cffi-lua implements; in luajit
ffi.nullptr will just become nil and it'll work like before.

4 years agoelua: remove old lua bindings generator
Daniel Kolesa [Thu, 21 May 2020 15:29:23 +0000 (17:29 +0200)]
elua: remove old lua bindings generator

4 years agoelm_genlist/elm_gengrid: fix memory leak caused by efl_wref_del.
Hosang Kim [Thu, 21 May 2020 09:36:10 +0000 (18:36 +0900)]
elm_genlist/elm_gengrid: fix memory leak caused by efl_wref_del.

Summary:
efl_wref_del makes pointer to NULL.
So we should copy pointer before efl_wref_del is called.

Reviewers: SanghyeonLee, cedric

Reviewed By: SanghyeonLee

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11866

4 years agowqRevert "spec: restore omitted config file for elementary"
Jongmin Lee [Wed, 20 May 2020 19:53:16 +0000 (04:53 +0900)]
wqRevert "spec: restore omitted config file for elementary"

This reverts commit 9f0cdf9addaa493cf69c71d5de8feec854471f34.

4 years agoevas_font: resolve dereference after null check issue
Jongmin Lee [Mon, 18 May 2020 06:52:39 +0000 (15:52 +0900)]
evas_font: resolve dereference after null check issue

Change-Id: Ic4e7f73f69cbd06bbe9809b43752c69ce818499e

4 years agospec: restore omitted config file for elementary
Wonki Kim [Wed, 20 May 2020 01:14:45 +0000 (10:14 +0900)]
spec: restore omitted config file for elementary

elementary has a default config file that has omitted.
this patch restores it.

Change-Id: I9d945e109d7a5ec9507f6f4fa18f581183e8a227

4 years agospec: add checkme definition to spec file
Wonki Kim [Wed, 20 May 2020 01:14:34 +0000 (10:14 +0900)]
spec: add checkme definition to spec file

this patch add a definition for checkme file for ecore_con

Change-Id: Ib86b8a2dd46d110095365f1f30f5de0b961e5af1

4 years agocheckme: add a 'checkme' file for ecore_con
Wonki Kim [Wed, 20 May 2020 01:14:20 +0000 (10:14 +0900)]
checkme: add a 'checkme' file for ecore_con

the file for ecore_con has missed.

Change-Id: Ib2e53626ed049edfc0ee10fe691c819e014d3b7c

4 years agowin/widget: free array when object is deleted.
Woochanlee [Wed, 20 May 2020 04:42:32 +0000 (13:42 +0900)]
win/widget: free array when object is deleted.

It makes a memory leak.

Change-Id: I25cd5c742f8ebbf0ace4511af38505aac6bc892e

4 years agomeson: Changed meson build to support address sanitizer build
Subhransu Mohanty [Tue, 21 Jan 2020 02:06:17 +0000 (02:06 +0000)]
meson: Changed meson build to support address sanitizer build

As we run some intermediate binary like edje_cc during efl build , and due to
those binary has some memory leaks the address sanitizer build of efl fails.
This Patch runs those binary with 'ASAN_OPTIONS=detect_leak=false' to disable the
memory leak check for those binaries.

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11137

4 years agoefl_ui_layout: remove group_calcualte to respect legacy min size
Yeongjong Lee [Wed, 20 May 2020 07:36:15 +0000 (16:36 +0900)]
efl_ui_layout: remove group_calcualte to respect legacy min size

Summary:
For legacy elm_layout, the size should be calculated based on
`hint_size_combined_min` instead of `hint_size_min` in the group_calculate.

Also, `_efl_ui_layout_efl_canvas_group_group_calculate` is unnecessary because
it is added to allow finger size policy differentiation between inherited layout
and layout object (see also, 46cf288d32f06a0c82ef227c65e55062bf494672), however,
it can be set by `finger_size_multiplier` proeprty in constructor.

Test Plan:
Check `evas_object_size_hint_min_set` is respected.

layout_example.c
```
//Compile with:
//edje_cc layout_example.edc &&  gcc -g layout_example.c -o layout_example `pkg-config --cflags --libs elementary`
#include <Elementary.h>

EAPI_MAIN int
elm_main(int argc EINA_UNUSED, char **argv EINA_UNUSED)
{
   Evas_Object *win, *layout, *box;

   elm_policy_set(ELM_POLICY_QUIT, ELM_POLICY_QUIT_LAST_WINDOW_CLOSED);

   win = elm_win_util_standard_add("layout", "Layout");
   elm_win_autodel_set(win, EINA_TRUE);

   box = elm_box_add(win);
   evas_object_size_hint_weight_set(box, EVAS_HINT_EXPAND, EVAS_HINT_EXPAND);
   evas_object_size_hint_align_set(box, EVAS_HINT_FILL, EVAS_HINT_FILL);
   evas_object_show(box);
   elm_win_resize_object_add(win, box);

   layout = elm_layout_add(box);
   evas_object_size_hint_min_set(layout, 300, 200);
   elm_layout_file_set(layout, "./layout_example.edj", "mylayout");
   evas_object_show(layout);
   elm_box_pack_end(box, layout);

   evas_object_resize(win, 300, 300);
   evas_object_show(win);

   elm_run();

   return 0;
}
ELM_MAIN()
```

layout_example.edc
```
collections {
   group {
      name: "mylayout";
      parts {
         part {
            name: "bg";
            type: RECT;
            description {
               state: "default" 0.0;
               color: 100 190 100 255;
            }
         }
         part {
            name: "mytext1";
            type: TEXT;
            description {
               state: "default" 0.0;
               text {
                  size: 30;
                  text: "my text1";
                  min: 1 0;
               }
               color: 0 255 0 255;
               rel1.relative: 0.0 0.0;
               rel2.relative: 1.0 0.3;
            }
         }
      }
      parts {
         part {
            name: "mytext2";
            type: TEXT;
            description {
               state: "default" 0.0;
               text {
                  size: 30;
                  text: "my text2";
                  min: 1 0;
               }
               color: 0 255 255 255;
               rel1.relative: 0.0 0.3;
               rel2.relative: 1.0 0.6;
            }
         }
      }
   }
}
```

Reviewers: zmike, Jaehyun_Cho, woohyun

Reviewed By: Jaehyun_Cho

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11858

4 years agogenlist: apply l10n translation for langauge change
SangHyeon Jade Lee [Mon, 18 May 2020 09:52:16 +0000 (18:52 +0900)]
genlist: apply l10n translation for langauge change

Change-Id: Ia2cf7fb3484a7157bfb47aa00d6e04fc6728d4d6

4 years agoecore-x - ensure we fully look at all xkb events and refresh all binds
Carsten Haitzler (Rasterman) [Tue, 19 May 2020 17:41:29 +0000 (18:41 +0100)]
ecore-x - ensure we fully look at all xkb events and refresh all binds

this ensures for sure that we rebind known key/mouse grabs too for xkb
mapping changes and force a get of the keymaps too

@fix

4 years agoevas: gesture_manager: fix usage of long on Windows
Vincent Torri [Fri, 15 May 2020 03:42:30 +0000 (03:42 +0000)]
evas: gesture_manager: fix usage of long on Windows

On Windows 64bits, long is a 4 bytes type. Declare count as being a intptr_t instead of long

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11836

4 years agotests: eina: fix usage of long on Windows
Vincent Torri [Fri, 15 May 2020 03:49:17 +0000 (03:49 +0000)]
tests: eina: fix usage of long on Windows

On Windows 64bits, long is a 4 bytes type. Declare i as being a uintptr_t instead of unsigned long

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Reviewed-by: Felipe Magno de Almeida <felipe@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11837

4 years agoecore_con: Fix unused argument warning
Vincent Torri [Tue, 19 May 2020 04:21:34 +0000 (04:21 +0000)]
ecore_con: Fix unused argument warning

Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11849

4 years agoevas: do not call evas_object_change in reneder_post
Shinwoo Kim [Tue, 19 May 2020 04:14:29 +0000 (13:14 +0900)]
evas: do not call evas_object_change in reneder_post

Summary:
evas_render_updates_internal > eina_array_remove is calling pending_change
pending_change > render_post could call evas_object_change >
evas_render_object_recalc.

The eina_array is removing its item from pending_objects, while
the evas_object_change that is calling evas_render_object_recalc is adding
item to pending_objects.

As a result, the pending_objects.count is incorrect, and it breaks render.

Reviewers: Hermet, jsuya, herb

Reviewed By: Hermet

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11835

4 years agoecore_anim: rework bezier curve function
Wonki Kim [Tue, 19 May 2020 04:05:11 +0000 (13:05 +0900)]
ecore_anim: rework bezier curve function

Summary:
current cubic bezier function isn't accurate at sometime.
to make it more accurate, this patch rework bezier curve
by using a cardano's algorithm.
(refer to https://pomax.github.io/bezierinfo/)

Reviewers: Hermet, bu5hm4n

Reviewed By: Hermet

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11819

4 years agopng: handle file name .9.png
Shinwoo Kim [Tue, 19 May 2020 02:57:55 +0000 (11:57 +0900)]
png: handle file name .9.png

Summary: .9.png is not 9 patch file, but a png file which name is .9

Test Plan: evas_object_image_file_set(image, "./.9.png", NULL);

Reviewers: Hermet, jsuya, herb

Reviewed By: Hermet

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11848

4 years agoelput: Rmove empty pub_eo_file_target in meson.build
Christopher Michael [Mon, 18 May 2020 15:48:08 +0000 (11:48 -0400)]
elput: Rmove empty pub_eo_file_target in meson.build

4 years agoelput: Fix formatting
Christopher Michael [Mon, 18 May 2020 15:26:34 +0000 (11:26 -0400)]
elput: Fix formatting

NB: No functional changes

4 years agoecore-evas-drm: Remove duplicate config set in meson.build file
Christopher Michael [Mon, 18 May 2020 15:24:48 +0000 (11:24 -0400)]
ecore-evas-drm: Remove duplicate config set in meson.build file

4 years agoelgind - it's apparently libelogind-shared ...
Carsten Haitzler (Rasterman) [Mon, 18 May 2020 10:13:22 +0000 (11:13 +0100)]
elgind - it's apparently libelogind-shared ...

4 years agoecore - systemd - missed an ifdef for sytemd disable
Carsten Haitzler (Rasterman) [Mon, 18 May 2020 09:27:55 +0000 (10:27 +0100)]
ecore - systemd - missed an ifdef for sytemd disable

4 years agosystemd - make libsystemd use/supprot entirely runtime "dlopened"
Carsten Haitzler (Rasterman) [Fri, 15 May 2020 11:24:36 +0000 (12:24 +0100)]
systemd - make libsystemd use/supprot entirely runtime "dlopened"

so i've moved all systemd and elogind support to be runtime only with
dlopen (eina_module) of libsystemd.so.0 (or libelogind.so.0 for elput)
and finding of symbols manually at runtime (if the right code paths or
env vars are set), thus remvoing the need to decide at compile time if
efl needs systemd support or not as it no longer needs systemd
headers/libs at compile time and just at runtime. this simplifies
building a bit and makes efl more adaptive to the final target system
at runtime.

4 years agoefl_ui_image: fix to call "clicked" smart callback when there is no img
WooHyun Jung [Mon, 18 May 2020 03:39:59 +0000 (12:39 +0900)]
efl_ui_image: fix to call "clicked" smart callback when there is no img

Summary:
Some applications have used "clicked" smart callback from elm_image which
does not have an image in it.
To keep backward compatibility, hit_rect needs to be resized properly to
call smart callback function.

@fix

Reviewers: Hermet, jsuya, kimcinoo

Reviewed By: Hermet

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11840

4 years agoAccessibility: __elm_atspi_bridge_highlighted_object_read API update
Shilpa Singh [Sun, 10 May 2020 20:59:56 +0000 (02:29 +0530)]
Accessibility: __elm_atspi_bridge_highlighted_object_read API update

__elm_atspi_bridge_highlighted_object_read API to be made as public

Change-Id: I693e7af2ba9a998a78761b01f79f42eb17eee558
Signed-off-by: Shilpa Singh <shilpa.singh@samsung.com>
4 years agoevas font: add/apply font glyph lru list
Youngbok Shin [Sat, 24 Mar 2018 04:58:44 +0000 (13:58 +0900)]
evas font: add/apply font glyph lru list

Nowadays, colored emoji fonts can have above 2000 bitmap glyphs.
If an application want to show whole emoji list, it will take
huge memory spaces.
For example, an emoji font have 2600 emoji glyphs. And each glyph
has 108x108px images. We can roughly calculate memory usages.
 108 * 108 * 4 * 2600 = 121,305,600 bytes = (roughly) 115 MB

So, Evas needs a feature which trims unused glyphs. Not unused fonts.
(It is all about one font.)

This patch adds glyph lru list.
When a glyph is added to/deleted from an array of glyphs, glyphs ref count will updated.
When a glyph's ref count become 0, it will be added to glyph lru list.
And Evas will try to trim whenever contents of Text_Props is free'd.

It is being reviewed in the following link.
https://phab.enlightenment.org/D5841

@tizen_feature

Change-Id: Ia4a2ddd6eab637b1afb1a56e3940a72525d7c98c

4 years agoRevert "evas_font_query: fixed the wrong arrangement when white-char is on wrapping...
WooHyun Jung [Fri, 15 May 2020 03:35:58 +0000 (12:35 +0900)]
Revert "evas_font_query: fixed the wrong arrangement when white-char is on wrapping point"

This reverts commit 9f50447c1dbbc67f05add96fad12c86a282d0ef0.

Change-Id: Iaf713d5584161a446b83793a7048e478891152a1

4 years agoevas_object_textblock: fixing white-space based wrapping + ellipsis issue.
WooHyun Jung [Fri, 15 May 2020 03:30:43 +0000 (12:30 +0900)]
evas_object_textblock: fixing white-space based wrapping + ellipsis issue.

1. remove _layout_line_reorder
  - This reodering should not be called becuase _layout_handle_ellipsis is
    not considering this reordering.

2. revome visually-deleted item from line items if it's added before

@tizen_fix

Change-Id: I5fa39f6ef394ba20bc7c7f16887269dde3f301c8

4 years agoecore_main: add more ttrace tracking point
Wonki Kim [Fri, 15 May 2020 04:14:04 +0000 (13:14 +0900)]
ecore_main: add more ttrace tracking point

this patch adds more ttrace tracking point where loop sleeps

Change-Id: I4455b9299fc4a3d35fc5d5488900b0cbdfea8544

4 years agoFix wrong tracking point
Seungha Son [Fri, 15 May 2020 03:19:49 +0000 (12:19 +0900)]
Fix wrong tracking point

Change-Id: I51cee369f6416460fb95d669a463b1cfac55a351
Signed-off-by: Seungha Son <seungha.son@samsung.com>
4 years agoecore-drm2: Add missing params for doxygen
Christopher Michael [Fri, 15 May 2020 14:50:37 +0000 (10:50 -0400)]
ecore-drm2: Add missing params for doxygen

4 years agoecore-drm2: Add API function to return default depth
Christopher Michael [Thu, 7 May 2020 15:28:05 +0000 (11:28 -0400)]
ecore-drm2: Add API function to return default depth

Rather than hard-coding depth & bpp in the evas drm engines, we can
use this function to return the default depth & bpp.

@feature

4 years agoevas_vg_shape: return NULL in evas_vg_shape_add when the parent of shape is NULL
Taehyub Kim [Fri, 15 May 2020 09:09:51 +0000 (18:09 +0900)]
evas_vg_shape: return NULL in evas_vg_shape_add when the parent of shape is NULL

Summary:
evas_vg_shaep_add should returned NULL when the parent is NULL
because the function does not allow the NULL parent.
@fix

Reviewers: jsuya, Hermet

Reviewed By: Hermet

Subscribers: cedric, #reviewers, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D11839

4 years agoelc_popup: do not allow theme apply while constructing for performance
Taehyub Kim [Thu, 14 May 2020 05:29:59 +0000 (14:29 +0900)]
elc_popup: do not allow theme apply while constructing for performance

Change-Id: I4df71eade988fd98159ea902124ece7466e57ce5