platform/upstream/efl.git
5 years agoevas/main: call correct destroy function in evas_free
Mike Blumenkrantz [Thu, 5 Jul 2018 19:12:45 +0000 (15:12 -0400)]
evas/main: call correct destroy function in evas_free

Summary:
an evas may or may not have a parent; this is legacy api and it's all
confusing

Reviewers: bu5hm4n, devilhorns

Reviewed By: bu5hm4n

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6486

5 years agotests/eo: add explicit test case for setting an object as its own parent
Mike Blumenkrantz [Thu, 5 Jul 2018 19:11:06 +0000 (15:11 -0400)]
tests/eo: add explicit test case for setting an object as its own parent

Summary: Depends on D6489

Reviewers: bu5hm4n, devilhorns

Reviewed By: bu5hm4n

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6490

5 years agoeo/base_class: explicitly refuse to set an object as its own parent
Mike Blumenkrantz [Fri, 29 Jun 2018 21:10:26 +0000 (17:10 -0400)]
eo/base_class: explicitly refuse to set an object as its own parent

somehow this has been possible until now. no.

@fix

Differential Revision: https://phab.enlightenment.org/D6489

5 years agoelm_popup: ensure focus is emitted on this object
Marcel Hollerbach [Thu, 5 Jul 2018 19:09:21 +0000 (15:09 -0400)]
elm_popup: ensure focus is emitted on this object

Summary:
This mirrors the focus property from the notify to the one from popup.

This is needed to establish legacy behaviour, as earlier popup was able to get focus.

ref T6707
Depends on D6510

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6707

Differential Revision: https://phab.enlightenment.org/D6511

5 years agoefl_ui_focus_manager_root_focus: mirror the focus properties correctly
Marcel Hollerbach [Thu, 5 Jul 2018 19:09:18 +0000 (15:09 -0400)]
efl_ui_focus_manager_root_focus: mirror the focus properties correctly

Summary:
this object takes a internal dummy object, and thus the properties
should be mirrored to the real root.

The focus property on the root of the manager is required that the outer world really sees that the focus is on the root or not.

Depends on D6506

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6510

5 years agoelua: disable this for now
Marcel Hollerbach [Thu, 5 Jul 2018 19:09:14 +0000 (15:09 -0400)]
elua: disable this for now

Summary:
the files that are generated are calling files that dont exist anymore.
This cannot work and is only eating time when building, disable this for
now. It can be reenabled late.
The api missing was removed in fc88037977dcc39dfd6d817c522cce01f5bfa024.

Reviewers: devilhorns, ManMower

Reviewed By: ManMower

Subscribers: ManMower, cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6506

5 years agoelementary: Add missing EINA_UNUSED for unused function parameter
Chris Michael [Wed, 4 Jul 2018 20:51:21 +0000 (16:51 -0400)]
elementary: Add missing EINA_UNUSED for unused function parameter

5 years agoecore: fix public/private class separation
Xavi Artigas [Wed, 4 Jul 2018 20:46:26 +0000 (16:46 -0400)]
ecore: fix public/private class separation

Summary:
A few classes were added in the wrong place in the makefile, rendering them
neither public nor legacy, and unaccessible to bindings.

Test Plan:
The Efl.Model_Item class is now accessible to C#, and previously it wasn't.
This class is used in the EO tutorials, which would have needed a rewrite.
make check, examples & distcheck still work.

Reviewers: bu5hm4n, zmike, devilhorns

Reviewed By: zmike, devilhorns

Subscribers: cedric, felipealmeida, lauromoura, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6507

5 years agoMigration: Fix Build Break
Shinwoo Kim [Thu, 19 Jul 2018 09:25:53 +0000 (18:25 +0900)]
Migration: Fix Build Break

5 years ago[Title] Fix to the bug which is the shared evas_gl_context with multi window
Jiyoun Park [Thu, 24 May 2018 08:59:33 +0000 (17:59 +0900)]
[Title]    Fix to the bug which is the shared evas_gl_context with multi window

[Problem]  When multi windows are shown, the latest window does not shown.
[Cause]    The latest window can not call glViewport with the updated coordinate
[Solution] Force call glViewport when window resizing occurs
[Team]     Graphics

Change-Id: I6df3f021385fe24190854abe84cb180d236a9dda

5 years agoemile image: close file properly.
Hermet Park [Wed, 18 Jul 2018 05:26:16 +0000 (14:26 +0900)]
emile image: close file properly.

@fix

Change-Id: Ied6436f4bc4cbb048a5c136ee000909a8bbd3aeb

5 years agoecore_evas_tbm : add tbm_surface_map and unmap at tbm_pixels_acquire()
Joogab Yun [Tue, 17 Jul 2018 08:15:29 +0000 (17:15 +0900)]
ecore_evas_tbm : add tbm_surface_map and unmap at tbm_pixels_acquire()
and release()

Change-Id: I6522c3d7730591cce7975b9d15d6ebc066e4174d

5 years agoevas : modified to share resource by each gl backend
Joogab Yun [Tue, 17 Jul 2018 07:53:33 +0000 (16:53 +0900)]
evas : modified to share resource by each gl backend

Change-Id: I5dfaa8ea94a840f3e1512707699bda8b8e2abf19

5 years agoevas_tbm : Add more thread safety
Joogab Yun [Mon, 2 Jul 2018 02:46:00 +0000 (11:46 +0900)]
evas_tbm : Add more thread safety

Change-Id: I093469f2f081205f9fa633487313c5b9bded4521

5 years agoevas : Change tbm_bufmgr to tbm_dummy_display.
Joogab Yun [Mon, 18 Jun 2018 04:32:14 +0000 (13:32 +0900)]
evas : Change tbm_bufmgr to tbm_dummy_display.

Change-Id: I7d9d88f01a6ff9960b6e0149ee0a55c2e80b41e3

5 years agoRevert "Revert "evas_gl: fix a TC fail of EGL_SYNC""
Hermet Park [Fri, 13 Jul 2018 05:11:01 +0000 (05:11 +0000)]
Revert "Revert "evas_gl: fix a TC fail of EGL_SYNC""

This reverts commit e491817c99e5a5d909af8b11453523dcede495ee.

Change-Id: I55bd46ebbfbff5390a88fa7c09de042fff2ac436

5 years agoRevert "Revert "evas : remove return null""
Hermet Park [Fri, 13 Jul 2018 05:10:44 +0000 (05:10 +0000)]
Revert "Revert "evas : remove return null""

This reverts commit 0c6502556008708c8441784525e2015c521c91d8.

Change-Id: I85b4f7e9cf9b3401336e634f7d52e572d8ea80fb

5 years agoRevert "evas : remove return null"
Hermet Park [Fri, 13 Jul 2018 05:08:26 +0000 (14:08 +0900)]
Revert "evas : remove return null"

This reverts commit 87eba6aa9eb2d31f496f845df4c3a79d73405ec1.

temporarily revert for SR.

Change-Id: I70deb73fba407212d585b01dc2cff728a328f0c1

5 years agoRevert "evas_gl: fix a TC fail of EGL_SYNC"
Hermet Park [Fri, 13 Jul 2018 05:07:38 +0000 (14:07 +0900)]
Revert "evas_gl: fix a TC fail of EGL_SYNC"

This reverts commit a757cd4e390f3f01261f1d3179632be0588354bc.

temporarily revert for SR.

Change-Id: I26d6feb0cae96f38e266df48e66c88f17776c576

5 years agoevas_gl: fix a TC fail of EGL_SYNC
Daekwang Ryu [Thu, 12 Jul 2018 06:02:03 +0000 (15:02 +0900)]
evas_gl: fix a TC fail of EGL_SYNC

Change-Id: I55e5972f82b33fc5af155cb85d6e7550f47f08ca

5 years agoFix crash in tct tests
Radoslaw Cybulski [Thu, 12 Jul 2018 09:35:19 +0000 (11:35 +0200)]
Fix crash in tct tests

Change-Id: I5148a0791288b195ee67f6227ffb0118fd615799

5 years agoevas : remove return null
Joogab Yun [Wed, 11 Jul 2018 05:24:32 +0000 (14:24 +0900)]
evas : remove return null

Change-Id: I539c597e91da10f9e2d5efcf76b61ac68fccba41

5 years agoFix for invalid navigation on list widget
Radoslaw Cybulski [Tue, 10 Jul 2018 10:43:27 +0000 (12:43 +0200)]
Fix for invalid navigation on list widget

Change-Id: I0bc8b2855c44fb8075651e993d0f3c5e4869ec3c

5 years agoevas_gl: fix EVAS_GL_EGL_SYNC_ON
Daekwang Ryu [Mon, 25 Jun 2018 04:05:57 +0000 (13:05 +0900)]
evas_gl: fix EVAS_GL_EGL_SYNC_ON

Change-Id: I6ae26b9fff0db48a4ce001fd006d613b83b11c6a

5 years agoevas textblock: adds missing legacy types
Youngbok Shin [Mon, 9 Jul 2018 00:57:58 +0000 (09:57 +0900)]
evas textblock: adds missing legacy types

*upstream patch(need a review)
 https://phab.enlightenment.org/D6529

@tizen_fix

Change-Id: I5360694dcad68003c0444c1643d81e92e1f40e67

5 years agoelm_toolbar: fix flickering issue when item resized.
Bowon Ryu [Mon, 9 Jul 2018 05:24:05 +0000 (14:24 +0900)]
elm_toolbar: fix flickering issue when item resized.

resize the box first to avoid flickering of
saved state of the toolbar when ratation changed but not in case of scroll.

Change-Id: I1ba904a8c9a346f707b869f03ebd15e07284566c
Signed-off-by: Bowon Ryu <bowon.ryu@samsung.com>
5 years agogengrid: fix component_highlight_grab
Tomasz Jeschke [Mon, 25 Jun 2018 14:46:22 +0000 (16:46 +0200)]
gengrid: fix component_highlight_grab

Change-Id: I759fcadd0f0403ae5898c77840857c41780a1aae

5 years agoFix navigation at CtxPopup
Oskar Chodowicz [Thu, 5 Jul 2018 15:58:06 +0000 (17:58 +0200)]
Fix navigation at CtxPopup

Change-Id: I20604a51f69fb10e064435fd52bbb1a7e36ee34c

5 years agoFix of invalid scrolling from first to last
Oskar Chodowicz [Wed, 4 Jul 2018 12:48:58 +0000 (14:48 +0200)]
Fix of invalid scrolling from first to last

fix scroll from first to last at apps-screen

Change-Id: I3926b8302ed28924d232811213ff27ac3a4dcbfd

5 years agoevas : fixed downscale sampling issue
Joogab Yun [Thu, 5 Jul 2018 05:33:57 +0000 (14:33 +0900)]
evas : fixed downscale sampling issue

    Fixed an issue where image wave due to differences in sampling
    coordinates when overlapping images

Change-Id: Ice000379c0397b52fd3bd44c84e8a7bb517df6bc

5 years agomultibuttonentry: Add Tizen Only features to maintain compatibility.
Woochan Lee [Mon, 9 Jul 2018 10:15:15 +0000 (19:15 +0900)]
multibuttonentry: Add Tizen Only features to maintain compatibility.

Change-Id: I64684105f79a6d06a473ff4d889d1f706311f054

5 years agoRevert "Fix structure of AT-SPI genlist subtree"
Bowon Ryu [Thu, 5 Jul 2018 10:31:30 +0000 (19:31 +0900)]
Revert "Fix structure of AT-SPI genlist subtree"

This reverts commit e17303010bfe5fe53127c708fc7fa56fc346630b.

5 years agoRevert "[evasgl] change swap mode when rect size not match tile size"
Hermet Park [Wed, 4 Jul 2018 01:36:04 +0000 (10:36 +0900)]
Revert "[evasgl] change swap mode when rect size not match tile size"

This reverts commit 9a7a006c5e38e5991ed7707e7fb35e8ceb1a3839.

This stopped partial rendering that breaks tv performance.
We need to make sure the precise scenario that requires this patch.
There may be a better solution to clip the drawing region
 with exact multiple of tile size. see...

Change-Id: If9a792d9d3ed93c5fe92e9862f29f4345f3571bd

5 years agoRevert "eo: Turn thread-validation off for unstable Tizen 5.0 apps which are violatin...
Hermet Park [Tue, 3 Jul 2018 03:26:59 +0000 (12:26 +0900)]
Revert "eo: Turn thread-validation off for unstable Tizen 5.0 apps which are violating thread-safety."

This reverts commit 017b38c5e98bfa2906782ed172d5d9c997d39414.

No more requirement by VD, so forget this.

Change-Id: Icda4cc71147f0a3cae0c85666fc83243b7186469

5 years agoFix structure of AT-SPI genlist subtree
Tomasz Jeschke [Tue, 19 Jun 2018 10:51:51 +0000 (12:51 +0200)]
Fix structure of AT-SPI genlist subtree

Change-Id: I3eefa7a6ebcd2f69004032c1d2b706f8545365cc

6 years agoupdate res folder that has been changed in upstream.
Hermet Park [Wed, 4 Jul 2018 09:02:00 +0000 (18:02 +0900)]
update res folder that has been changed in upstream.

Change-Id: I12efe0ae0fcad4fdf65fbb96cf49cdb843012493

6 years agoeina: Improve eina_binbuf function documentation
Bryce Harrington [Wed, 4 Jul 2018 02:12:57 +0000 (11:12 +0900)]
eina: Improve eina_binbuf function documentation

Summary:
Define return values as part of @return.
Cleanup grammar.

Reviewers: devilhorns

Subscribers: cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6503

6 years agoevas: add more error messages when passing invalid objects to evas callbacks
Mike Blumenkrantz [Wed, 4 Jul 2018 01:51:01 +0000 (10:51 +0900)]
evas: add more error messages when passing invalid objects to evas callbacks

Summary:
this should only happen if the user has made a mistake regarding the
existence or type of an object, so ensure that an error message occurs to
help debug any failures which result

fix T6326

Reviewers: bu5hm4n, Hermet, woohyun, devilhorns

Reviewed By: Hermet

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6326

Differential Revision: https://phab.enlightenment.org/D6322

6 years agoeina: Improve eina_benchmark function documentation
Bryce Harrington [Tue, 3 Jul 2018 18:44:41 +0000 (14:44 -0400)]
eina: Improve eina_benchmark function documentation

Summary:
Define return values as part of @return.
Clarify that the returned list of files are gnuplot filenames specifically.
Cleanup grammar throughout.

Reviewers: zmike, devilhorns

Reviewed By: devilhorns

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6502

6 years agoefreet: use correct path when generating error message about stale efreetd socket
Mike Blumenkrantz [Tue, 3 Jul 2018 16:49:08 +0000 (12:49 -0400)]
efreet: use correct path when generating error message about stale efreetd socket

Summary:
using runtime directory in all cases for this is wrong, as ecore-con has a number
of fallback codepaths for the case where runtime directory is not set or not valid.

by using the same ecore-con function which ecore-ipc uses to generate the socket
string, the error message path should always be the same as the path which is
used by efreetd

extra linkage was required by efreet in order to use ecore-con functions, so
the internal lib variable in the build system was modified to provide this

@fix

fix T7045

Reviewers: devilhorns

Reviewed By: devilhorns

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T7045

Differential Revision: https://phab.enlightenment.org/D6425

6 years agotests/eina: disable eina debug signal handling test in non-fork mode
Mike Blumenkrantz [Tue, 3 Jul 2018 16:46:49 +0000 (12:46 -0400)]
tests/eina: disable eina debug signal handling test in non-fork mode

Summary: signal tests can only be used in fork mode or else they just fail

Reviewers: devilhorns

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6477

6 years agotests/ecore-con: fix dns test to use correct ordering
Mike Blumenkrantz [Tue, 3 Jul 2018 16:44:13 +0000 (12:44 -0400)]
tests/ecore-con: fix dns test to use correct ordering

Summary:
it's unclear to me why this test was implemented with the expected
event ordering of SERVER_DEL -> SERVER_ERROR; going as far back
as efl 1.7, the behavior has always been ERROR -> DEL

this fixes the test to verify the expected event ordering on a dns
resolve failure
Depends on D6481

Reviewers: devilhorns

Reviewed By: devilhorns

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6482

6 years agoecore-con/legacy: kill server with DEL event if a dialer error occurs
Mike Blumenkrantz [Tue, 3 Jul 2018 16:42:00 +0000 (12:42 -0400)]
ecore-con/legacy: kill server with DEL event if a dialer error occurs

Summary:
in the course of the efl-net rewrite, the previous relied-upon behavior
sequence of error -> disconnect -> del was broken, and error events no
longer triggered disconnects or server deletion.

the failure was caused by a check for the 'connecting' flag, preventing
DEL events from being triggered

further, the ordering of the events was flipped from ERROR -> DEL to
DEL -> ERROR; this is also fixed now

this caused any failed connection to leak the entire server connection
since existing handlers for the del event were never triggered and the
server was never destroyed

@fix
fix T6330

Reviewers: devilhorns

Reviewed By: devilhorns

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6330

Differential Revision: https://phab.enlightenment.org/D6481

6 years agoelm_test: fix invalid read in dnd image drop
Mike Blumenkrantz [Tue, 3 Jul 2018 16:39:36 +0000 (12:39 -0400)]
elm_test: fix invalid read in dnd image drop

Summary:
this selection is data with a specified length, not a terminated string

@fix
Depends on D6484

Reviewers: devilhorns

Subscribers: netstar, cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6485

6 years agoefl/selection_manager: fix some invalid memory reads
Mike Blumenkrantz [Tue, 3 Jul 2018 16:39:33 +0000 (12:39 -0400)]
efl/selection_manager: fix some invalid memory reads

Summary:
these selections are data with specified lengths, not strings
Depends on D6483

Reviewers: devilhorns, bu5hm4n

Subscribers: bu5hm4n, netstar, cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6484

6 years agoefl/selection_manager: make selection manager a child of the app
Mike Blumenkrantz [Tue, 3 Jul 2018 16:39:26 +0000 (12:39 -0400)]
efl/selection_manager: make selection manager a child of the app

Summary:
a selection manager is application-wide, not per-window. creating separate
managers for each window duplicates all callbacks for the window's display
server, guaranteeing broken behavior at any time when more than one window
exists

fix T6937

Reviewers: bu5hm4n, devilhorns

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6937

Differential Revision: https://phab.enlightenment.org/D6483

6 years agoeina/debug: remove global lock for session list access
Mike Blumenkrantz [Mon, 2 Jul 2018 18:28:35 +0000 (14:28 -0400)]
eina/debug: remove global lock for session list access

this list is now only accessed from the main thread so there's no reason
to have a lock for it

6 years agoeina/debug: rework session thread destruction to avoid deadlock on shutdown
Mike Blumenkrantz [Mon, 2 Jul 2018 18:14:32 +0000 (14:14 -0400)]
eina/debug: rework session thread destruction to avoid deadlock on shutdown

when debugging was active, this would frequently result in the monitor thread
blocking in a read() call for a fd which had already been closed by the main
thread. the main thread would be waiting for the monitor thread to fail its read()
and destroy itself, but the read() would never end without an interrupt

this reworks the monitor thread destruction to stop relying on the thread to
remove its session from the session list and instead only access the list from
the main thread, allowing sessions to always be immediately destroyed

fix T7103

6 years agoefl_ui_focus_layer: do not call on NULL objects
Marcel Hollerbach [Sat, 30 Jun 2018 10:37:58 +0000 (12:37 +0200)]
efl_ui_focus_layer: do not call on NULL objects

This fixes the following backtrace

#5  0x00007ffff4ee87d0 in efl_ui_focus_manager_root_get (obj=0x0) at lib/elementary/efl_ui_focus_manager.eo.c:17
#6  0x00007ffff4ef50cb in _efl_ui_focus_layer_enable_set (obj=0x40000004adfe, pd=0x12fde40, v=0 '\000') at lib/elementary/efl_ui_focus_layer.c:127
#7  0x00007ffff4ef52dd in efl_ui_focus_layer_enable_set (obj=0x40000004adfe, v=1 '\001') at lib/elementary/efl_ui_focus_layer.eo.c:4
#8  0x00007ffff4ef4df0 in _efl_ui_focus_layer_efl_gfx_entity_visible_set (obj=0x40000004adfe, pd=0x12fde40, v=1 '\001') at lib/elementary/efl_ui_focus_layer.c:35
#9  0x00007ffff0f366a3 in efl_gfx_entity_visible_set (obj=0x40000004adfe, v=1 '\001') at ../src/lib/efl/interfaces/efl_gfx_entity.eo.c:15
#10 0x00007ffff5d832bb in evas_object_show (eo_obj=0x40000004adfe) at lib/evas/canvas/evas_object_main.c:1917
#11 0x00007fffcc365c74 in _bar_icon_preview_show (data=0xfaa7e0) at ../src/modules/luncher/bar.c:906
#12 0x00007ffff6d4305a in _ecore_call_task_cb (func=0x7fffcc36549c <_bar_icon_preview_show>, data=0xfaa7e0) at lib/ecore/ecore_private.h:442
#13 0x00007ffff6d434d2 in _ecore_timer_legacy_tick (data=0x12ca950, event=0x7ffffffc57c0) at lib/ecore/ecore_timer.c:160
#14 0x00007ffff0d06e90 in _event_callback_call (obj_id=0x4000001f25b2, pd=0x923820, desc=0x7ffff6f673e0 <_EFL_LOOP_TIMER_EVENT_TICK>, event_info=0x0, legacy_compare=0 '\000') at lib/eo/eo_base_class.c:1671
#15 0x00007ffff0d07162 in _efl_object_event_callback_call (obj_id=0x4000001f25b2, pd=0x923820, desc=0x7ffff6f673e0 <_EFL_LOOP_TIMER_EVENT_TICK>, event_info=0x0) at lib/eo/eo_base_class.c:1755
#16 0x00007ffff0d07208 in efl_event_callback_call (obj=0x4000001f25b2, desc=0x7ffff6f673e0 <_EFL_LOOP_TIMER_EVENT_TICK>, event_info=0x0) at lib/eo/eo_base_class.c:1758
#17 0x00007ffff6d4478e in _efl_loop_timer_expired_call (obj=0x4000000003dc, pd=0x890bf0, when=333436.894054887) at lib/ecore/ecore_timer.c:634
#18 0x00007ffff6d445bd in _efl_loop_timer_expired_timers_call (obj=0x4000000003dc, pd=0x890bf0, when=333436.894054887) at lib/ecore/ecore_timer.c:587
#19 0x00007ffff6d0b423 in _ecore_main_loop_iterate_internal (obj=0x4000000003dc, pd=0x890bf0, once_only=0) at lib/ecore/ecore_main.c:2317
#20 0x00007ffff6d08e66 in _ecore_main_loop_begin (obj=0x4000000003dc, pd=0x890bf0) at lib/ecore/ecore_main.c:1175
#21 0x00007ffff6d11757 in _efl_loop_begin (obj=0x4000000003dc, pd=0x890bf0) at lib/ecore/efl_loop.c:83
#22 0x00007ffff6d13e6e in efl_loop_begin (obj=0x4000000003dc) at lib/ecore/efl_loop.eo.c:28
#23 0x00007ffff6d08fe0 in ecore_main_loop_begin () at lib/ecore/ecore_main.c:1248
#24 0x000000000054817c in main (argc=2, argv=0x7fffffffdc78) at ../src/bin/e_main.c:1090

fix T7034.

Differential Revision: https://phab.enlightenment.org/D6492

6 years agoelm_toolbar: handle the cleanup in the invalidate call
Marcel Hollerbach [Mon, 2 Jul 2018 09:16:41 +0000 (11:16 +0200)]
elm_toolbar: handle the cleanup in the invalidate call

there is actaully no need to add the event, we can cleanup the pointer
in the invalidate call. Further more, the event was executed before the
method, thus checks against more_item failed because that was already
NULL. The result have been error messages, this fixes the error.

Differential Revision: https://phab.enlightenment.org/D6496

6 years agoelm_toolbar: fix _item_visiblity
Marcel Hollerbach [Mon, 2 Jul 2018 08:58:52 +0000 (10:58 +0200)]
elm_toolbar: fix _item_visiblity

'more' should indicate when a more item is needed, this is needed if one
item is set to visibility false, thus checking this here is required.

 #FocusBug

fix T6806

Differential Revision: https://phab.enlightenment.org/D6495

6 years agoefl_ui_bg: set scale manually to hand by hand
Marcel Hollerbach [Mon, 2 Jul 2018 06:31:54 +0000 (08:31 +0200)]
efl_ui_bg: set scale manually to hand by hand

according to 09d3d5b85ad4a0b7441b5b1e508db0e595bef395 elm scale should
be ignored. However, setting no_scale breaks efl_ui_bg as no_scale
disables scale types, which are required to work.

fix T7080

Differential Revision: https://phab.enlightenment.org/D6494

6 years agoefl_ui_image: apply the whole sizing eval
Marcel Hollerbach [Mon, 2 Jul 2018 06:26:01 +0000 (08:26 +0200)]
efl_ui_image: apply the whole sizing eval

orientation no_scale etc. can have impact on the min max calculation.
Thus that should be called on every property set.

ref T7080

Differential Revision: https://phab.enlightenment.org/D6493

6 years agoefl_ui_focus_graph: improve results based on lvls
Marcel Hollerbach [Sat, 30 Jun 2018 09:46:29 +0000 (11:46 +0200)]
efl_ui_focus_graph: improve results based on lvls

This improves the results from the graph calculations.
Everything is documented in the comments.

The genlist of focus 6 is placed in a way that the button
at the top is not in reach, before we only had lvl1 and lvl2
(as described in the code comment), thus the button was only
reachable via right not via top. lvl3 makes it accessable
via top AND right.

fix T7098

Differential Revision: https://phab.enlightenment.org/D6491

6 years agoEvas Textblock: Don't convert <br/> to <ps/> after changing text
Youngbok Shin [Mon, 2 Jul 2018 13:17:33 +0000 (16:17 +0300)]
Evas Textblock: Don't convert <br/> to <ps/> after changing text

Summary:
It was only happened when legacy newline is enabled. By default,
legacy newline is enabled. As I know, legacy newline option has
to change textblock's internal behavior. But, it shouldn't change
the given original text. It fixes T3399.

Test Plan: A Test case is included in Evas test suite.

Reviewers: Jaehyun_Cho, z-wony, tasn, woohyun, herdsman, Blackmole, devilhorns

Subscribers: #committers, zmike, raster, cedric, jpeg

Tags: #efl

Maniphest Tasks: T3399

Differential Revision: https://phab.enlightenment.org/D3874

6 years agoelementary: fix buffer overflow in efl/calendar widget
Xavi Artigas [Mon, 2 Jul 2018 13:45:55 +0000 (13:45 +0000)]
elementary: fix buffer overflow in efl/calendar widget

I missed this one in the previous part name refactor, where the efl
namespace was added.
Differential Revision: https://phab.enlightenment.org/D6498

6 years agodata: install icon according to FDO specification.
Alastair Poole [Mon, 2 Jul 2018 12:37:51 +0000 (12:37 +0000)]
data: install icon according to FDO specification.

install to icons/hicolor/128x128/apps instead of icons/
Differential Revision: https://phab.enlightenment.org/D6497

6 years agoefl_ui_focus_parent_provider_gen: provide results for adapters
Marcel Hollerbach [Mon, 2 Jul 2018 11:14:17 +0000 (20:14 +0900)]
efl_ui_focus_parent_provider_gen: provide results for adapters

Summary:
a adapter is used whenno real widget is created during realization, when
a adapter is used, the parent cannot be fetched via
elm_widget_parent_get, there we need efl_parent_get

Reviewers: ManMower, devilhorns, stephenmhouston, zmike, Hermet

Reviewed By: Hermet

Subscribers: Hermet, cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6339

6 years agoEvas textblock: clear ellipsis item at start of layout
Daniel Hirt [Fri, 29 Jun 2018 14:05:06 +0000 (17:05 +0300)]
Evas textblock: clear ellipsis item at start of layout

The ellipsis item was tracked so it can be inserted at the correct
position. It was not cleared properly, leading to a randomly failing
test suite.

Actually, it seems that there is no need to even store the ellipsis item
as the object data, so it was also moved to be in the layout context,
instead.

A special shoutout to ManMower for lending me his machine, as the test
suite didn't fail on mine.

Fixes T6986

@fix

6 years agoEvas font: simplify fallback parsing step
Daniel Hirt [Sat, 30 Jun 2018 07:15:00 +0000 (10:15 +0300)]
Evas font: simplify fallback parsing step

@fix

6 years agoEvas font: add comparison for fallbacks in evas_font_desc_cmp()
Youngbok Shin [Fri, 29 Jun 2018 20:09:08 +0000 (23:09 +0300)]
Evas font: add comparison for fallbacks in evas_font_desc_cmp()

Summary:
fallbacks string also has to be compared to load proper fdesc.
If a font which does not have fallback fonts is loaded,
fallback fonts can't be appended to the same font.
@fix

Reviewers: tasn, woohyun, herdsman, zmike, devilhorns

Reviewed By: herdsman

Subscribers: #committers, zmike, raster, Blackmole, z-wony, cedric, jpeg

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D3707

6 years agoEvas font: fix inifinite loop problem for fallback fonts
Youngbok Shin [Thu, 28 Jun 2018 15:50:50 +0000 (18:50 +0300)]
Evas font: fix inifinite loop problem for fallback fonts

Summary:
When multiple fallback fonts was passed to evas_font_load(),
the while loop could run forever. I think it is never tested.
@fix

Test Plan:
Set the following textblock style and set it to a textblock object.
"font=Sans font_fallbacks=Ubuntu,Inconsolata,BlahBlah font_size=..."

Run and see application is in stuck.

Reviewers: tasn, woohyun, herdsman, devilhorns

Subscribers: #committers, zmike, raster, Blackmole, z-wony, cedric, jpeg

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D3703

6 years agodefault theme: fix part names in elm/clock
Xavi Artigas [Fri, 29 Jun 2018 15:49:19 +0000 (11:49 -0400)]
default theme: fix part names in elm/clock

Summary: Ref T6965

Reviewers: devilhorns, zmike

Reviewed By: zmike

Subscribers: cedric, #committers, zmike

Tags: #efl

Maniphest Tasks: T6965

Differential Revision: https://phab.enlightenment.org/D6452

6 years agoefl_ui_focus_manager_calc: only perform operation if really required
Marcel Hollerbach [Fri, 29 Jun 2018 15:36:30 +0000 (11:36 -0400)]
efl_ui_focus_manager_calc: only perform operation if really required

Summary:
_manager_in_chain_set is not needed to be called if the focus property
is already correct. If a manager is moved out of the redirect property,
then the focus is dropped completly

Reviewers: ManMower, devilhorns, zmike, stephenmhouston

Reviewed By: zmike

Subscribers: cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6338

6 years agoeet: set parent struct member when adding a child to a parent node
Mike Blumenkrantz [Fri, 29 Jun 2018 15:36:12 +0000 (11:36 -0400)]
eet: set parent struct member when adding a child to a parent node

Summary:
this relation can be retrieved with eet_node_parent_get() but is never
set internally

fix T4600

@fix

Reviewers: artem.popov, devilhorns, Hermet, vtorri

Reviewed By: Hermet

Subscribers: Hermet, cedric, #committers

Tags: #efl

Maniphest Tasks: T4600

Differential Revision: https://phab.enlightenment.org/D6336

6 years agodefault theme: fix part names in efl/uiclock
Xavi Artigas [Fri, 29 Jun 2018 15:34:57 +0000 (11:34 -0400)]
default theme: fix part names in efl/uiclock

Summary:
Ref T6965
Depends on D6450

Reviewers: devilhorns, zmike

Reviewed By: zmike

Subscribers: cedric, #committers, zmike

Tags: #efl

Maniphest Tasks: T6965

Differential Revision: https://phab.enlightenment.org/D6451

6 years agodefault theme: remove unused elm/uiclock
Xavi Artigas [Fri, 29 Jun 2018 15:34:36 +0000 (11:34 -0400)]
default theme: remove unused elm/uiclock

Summary:
Legacy widget is elm/clock, and the new one is efl/uiclock.
There does not exist a legacy elm/uiclock.
This also reverts commit 20404d79d43876c546ba8e969f6cb9ea38122772
(elm_datetime, efl_ui_clock : Add check 'legacy widget' for layout signal emission)
Since there is no need to check for legacy versions of uiclock.

Ref T6965

Reviewers: devilhorns, zmike

Reviewed By: zmike

Subscribers: cedric, #committers, zmike

Tags: #efl

Maniphest Tasks: T6965

Differential Revision: https://phab.enlightenment.org/D6450

6 years agoelm_genlist: prepare items before realising them
Marcel Hollerbach [Fri, 29 Jun 2018 15:33:18 +0000 (11:33 -0400)]
elm_genlist: prepare items before realising them

Summary: Depends on D6339

Reviewers: ManMower, devilhorns, zmike, stephenmhouston

Reviewed By: zmike

Subscribers: Hermet, cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6340

6 years agodefault theme: fix part names in efl/calendar
Xavi Artigas [Fri, 29 Jun 2018 15:32:54 +0000 (11:32 -0400)]
default theme: fix part names in efl/calendar

Summary:
Ref T6967
Depends on D6453

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6967

Differential Revision: https://phab.enlightenment.org/D6454

6 years agodefault theme: fix part names in elm/calendar
Xavi Artigas [Fri, 29 Jun 2018 15:32:03 +0000 (11:32 -0400)]
default theme: fix part names in elm/calendar

Summary: Ref T6967

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6967

Differential Revision: https://phab.enlightenment.org/D6453

6 years agoefl_ui_focus_layer: elm_widget_top_get considered harmfull
Marcel Hollerbach [Fri, 29 Jun 2018 15:31:56 +0000 (11:31 -0400)]
efl_ui_focus_layer: elm_widget_top_get considered harmfull

Summary:
You think elm_widget_top_get returns a window object ? Oh no! For the
case that the widget tree splits, and the widget_parent gets NULL of
some widget in the chain, elm_widget_top_get returns where the NULL
reference is placed.

For the case of elm_notify this is somehow intended, as the notify does
not have a parent when the parent is deleted.

ref T7090

Reviewers: netstar, zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T7090

Differential Revision: https://phab.enlightenment.org/D6471

6 years agoeo: improve thread check during class construction
Mike Blumenkrantz [Fri, 29 Jun 2018 15:31:10 +0000 (11:31 -0400)]
eo: improve thread check during class construction

Summary:
a common use case is for a class to be constructed during a thread+mainloop
sync (e.g., ecore_thread_main_loop_begin() ecore_thread_main_loop_end())
and then naturally destroyed in the main thread during shutdown

ref 023a9ca2ee32529849e770f057f58592956dee47

Reviewers: bu5hm4n, devilhorns

Reviewed By: bu5hm4n

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6472

6 years agodefault theme: fix part names in notification
Xavi Artigas [Fri, 29 Jun 2018 15:29:56 +0000 (11:29 -0400)]
default theme: fix part names in notification

Summary:
Ref T6966
Depends on D6462

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966

Differential Revision: https://phab.enlightenment.org/D6463

6 years agodefault theme: fix part names in dialog
Xavi Artigas [Fri, 29 Jun 2018 15:29:40 +0000 (11:29 -0400)]
default theme: fix part names in dialog

Summary:
Ref T6966
Depends on D6461

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966

Differential Revision: https://phab.enlightenment.org/D6462

6 years agodefault theme: fix part names in illume
Xavi Artigas [Fri, 29 Jun 2018 15:29:33 +0000 (11:29 -0400)]
default theme: fix part names in illume

Summary:
Ref T6966
Depends on D6460

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966

Differential Revision: https://phab.enlightenment.org/D6461

6 years agodefault theme: fix part names in systray
Xavi Artigas [Fri, 29 Jun 2018 15:29:27 +0000 (11:29 -0400)]
default theme: fix part names in systray

Summary:
Ref T6966
Depends on D6459

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966

Differential Revision: https://phab.enlightenment.org/D6460

6 years agodefault theme: fix group names in music_control
Xavi Artigas [Fri, 29 Jun 2018 15:29:17 +0000 (11:29 -0400)]
default theme: fix group names in music_control

Summary:
Ref T7072
Depends on D6458

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966, T7072

Differential Revision: https://phab.enlightenment.org/D6459

6 years agodefault theme: fix group names in fileman
Xavi Artigas [Fri, 29 Jun 2018 15:29:08 +0000 (11:29 -0400)]
default theme: fix group names in fileman

Summary:
Ref T6966
Depends on D6457

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966

Differential Revision: https://phab.enlightenment.org/D6458

6 years agodefault theme: fix part names in elm/ews
Xavi Artigas [Fri, 29 Jun 2018 15:29:02 +0000 (11:29 -0400)]
default theme: fix part names in elm/ews

Summary:
Fixes T7046
Depends on D6456

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966, T7046

Differential Revision: https://phab.enlightenment.org/D6457

6 years agodefault theme: fix part names in elm/scrollable/panel
Xavi Artigas [Fri, 29 Jun 2018 15:28:57 +0000 (11:28 -0400)]
default theme: fix part names in elm/scrollable/panel

Summary:
Ref T6966

Depends on D6455

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966

Differential Revision: https://phab.enlightenment.org/D6456

6 years agodefault theme: fix part names in elm/popup
Xavi Artigas [Fri, 29 Jun 2018 15:28:50 +0000 (11:28 -0400)]
default theme: fix part names in elm/popup

Summary: Ref T6966

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T6966

Differential Revision: https://phab.enlightenment.org/D6455

6 years agoEvil: fix usage of Evil in a program compiled with vc++ this should be enough to...
Vincent Torri [Fri, 29 Jun 2018 13:56:15 +0000 (09:56 -0400)]
Evil: fix usage of Evil in a program compiled with vc++ this should be enough to fix T5206

Summary: Fix EFL errors when compiling a program with vc++

Test Plan: compilation

Reviewers: zmike, devilhorns, an.kroitor

Reviewed By: zmike

Subscribers: cedric, an.kroitor, zmike, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6478

6 years agoAdd information about the location of CC BY-SA 3.0 related icon files.
Myoungwoon Roy, Kim [Fri, 29 Jun 2018 09:57:19 +0000 (05:57 -0400)]
Add information about the location of CC BY-SA 3.0 related icon files.

Summary: COPYING.images does not have any information about the location of files.

Test Plan: N/A

Reviewers: cedric, jpeg, raster, Hermet, stefan_schmidt, Jaehyun_Cho, devilhorns

Reviewed By: devilhorns

Subscribers: #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6479

6 years agoconfigure: check if realpath program exists
Jaehyun Cho [Fri, 29 Jun 2018 05:43:03 +0000 (14:43 +0900)]
configure: check if realpath program exists

Summary: If realpath program does not exist, then readlink program can be used.

Test Plan:
Execute configure and check the log "checking for realpath".
If realpath program does not exist, then "checking for readlink" log is also printed.

Reviewers: woohyun, lauromoura, devilhorns

Reviewed By: lauromoura

Subscribers: cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6470

6 years agoElm label: fix "horizontal_fixed" usage with different styles
Daniel Hirt [Fri, 29 Jun 2018 01:25:57 +0000 (10:25 +0900)]
Elm label: fix "horizontal_fixed" usage with different styles

Summary:
This adds the "horizontal_fixed" description to each of the inheriting
parts in their respective inheriting groups "left" and "right".

An issue was observed when an emitted signal caused the part's
description to change to the one inherited from the "default" group.

Fixes T5382

@fix

Reviewers: #committers, devilhorns, Hermet

Reviewed By: #committers, Hermet

Subscribers: cedric, zmike

Tags: #efl

Maniphest Tasks: T5382

Differential Revision: https://phab.enlightenment.org/D6467

6 years agotests/eo: add explicit ref check when setting an invalid parent
Mike Blumenkrantz [Fri, 29 Jun 2018 01:21:01 +0000 (10:21 +0900)]
tests/eo: add explicit ref check when setting an invalid parent

Summary: ref f4a55e9266520773de96bfc55b603443cd61696d

Reviewers: bu5hm4n, devilhorns, Hermet

Reviewed By: Hermet

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6473

6 years agoevas map: check render condition more elaborately.
Hermet Park [Fri, 29 Jun 2018 00:58:41 +0000 (09:58 +0900)]
evas map: check render condition more elaborately.

Summary:
Ok, this was started from a bug that canvas getting not be updated.
If map is just disabled, at least one frame in the map region should be redrawn
So I added a condition 'map changed' in the render even though map is off
status. Now, I got a performance regression issue because it makes dirty
region is always true for the map object.

That is a corner case acutally, that object is not rendered but map still
have changed status.

I replaced the condition only if object is changed + map is changed.

At least, my test case works better with this patch.

@fix T6975

Reviewers: #committers, ManMower, devilhorns

Reviewed By: #committers, ManMower

Subscribers: ApB, ManMower, cedric, #committers, zmike

Tags: #efl

Maniphest Tasks: T6975

Differential Revision: https://phab.enlightenment.org/D6429

6 years agoevas: remove ecore init/shutdown calls from everywhere besides evas_main
Mike Blumenkrantz [Thu, 28 Jun 2018 20:03:15 +0000 (15:03 -0500)]
evas: remove ecore init/shutdown calls from everywhere besides evas_main

Summary:
these separate inits and shutdowns make it impossible to effectively control
ecore's lifetime which makes evas_shutdown unreliable as objects may be
destroyed at any point

ref T7052
Depends on D6475

Reviewers: ManMower, devilhorns

Reviewed By: ManMower, devilhorns

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T7052

Differential Revision: https://phab.enlightenment.org/D6476

Change-Id: I3d62667879896b0c3047ff7caebdc2f7f4c55dd7

6 years agoecore_evas: prevent duplicated ecore_evas registration 18/182518/2
YeongJong Lee [Mon, 25 Jun 2018 19:14:01 +0000 (15:14 -0400)]
ecore_evas: prevent duplicated ecore_evas registration

Summary:
After a44697c37a304c0d4d574b6772775c6936869e24, we can register same ecore_evas
to ecore_evases using ecore_evas_input_event_register.
(ecore_evas_input_event_register -> ecore_evas_done -> _ecore_evas_register)
This can make infinite loop in EINA_INLIST_FOREACH(ecore_evases, ee) because
next inlist of ecore_evases is ecore_evases after double call of
_ecore_evas_register.

This patch prevent it.

Test Plan:
Ecore_Evas *ee = ecore_evas_new(NULL, 0, 0, 800, 600, NULL);
ecore_evas_input_event_register(ee);
(part of document of ecore_fb_input_device_window_set)

Check that there is no infinite loop

Reviewers: zmike, devilhorns

Reviewed By: zmike

Subscribers: cedric, #committers

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6390

Change-Id: Ia1fc676abf2893849210c308e269e920b472784e

6 years agoRevert "ee_wayland_egl: Remove some amazing weirdness"
Jiyoun Park [Fri, 29 Jun 2018 04:04:00 +0000 (13:04 +0900)]
Revert "ee_wayland_egl: Remove some amazing weirdness"

calendar widget has black because of this patch

This reverts commit 0fdf4204a5977d66604f3a2bd65d76eb4ed4293e.

Change-Id: I9a9e0015b134965d1070f923ba9f699d730e7f37

6 years agoRevert "ecore_evas_wayland: Handle region updates and alpha setting in engine"
Jiyoun Park [Fri, 29 Jun 2018 04:03:26 +0000 (13:03 +0900)]
Revert "ecore_evas_wayland: Handle region updates and alpha setting in engine"

lock screen loses focus because of this patch

This reverts commit 26aa5dab423fd874bd9529d0682f9bc71f93819e.

6 years agoevas/main: move common shutdown to after ecore shutdown
Mike Blumenkrantz [Thu, 28 Jun 2018 19:53:10 +0000 (14:53 -0500)]
evas/main: move common shutdown to after ecore shutdown

Summary:
ecore_shutdown will trigger object deletions which require common
components to still be active in order to avoid crashes

ref 3433be343779424c5e030ace30e211298cd060f8
ref T7052

Reviewers: ManMower, devilhorns

Reviewed By: ManMower, devilhorns

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T7052

Differential Revision: https://phab.enlightenment.org/D6475

6 years agoecore_anim: Simplify timeline terminal condition
Derek Foreman [Thu, 28 Jun 2018 18:12:03 +0000 (13:12 -0500)]
ecore_anim: Simplify timeline terminal condition

Summary:
We should only ever have a pos of 1.0 once, the current terminal
condition gives the impression that it might be possible to have
more than one 1.0 firing.

This would break a lot of code.

No functional change intended.
Depends on D6464

Reviewers: devilhorns, zmike

Reviewed By: zmike

Subscribers: cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6465

6 years agoevas_render: Make some code easier to read
Derek Foreman [Thu, 28 Jun 2018 18:11:59 +0000 (13:11 -0500)]
evas_render: Make some code easier to read

Summary:
The variable "clean_them" can only ever be EINA_FALSE for much of this
function, but using it as a return value ensures that anyone not
intimately familiar with the code will have to read a lot of code
to figure out that this is so.

Instead, return EINA_FALSE up until the point clean_them can actually
be something else.

No functional change.

Reviewers: devilhorns, zmike

Reviewed By: zmike

Subscribers: cedric, #committers, zmike

Tags: #efl

Differential Revision: https://phab.enlightenment.org/D6464

6 years agoevas/main: shuffle shutdown order of filters and modules
Mike Blumenkrantz [Thu, 28 Jun 2018 16:41:12 +0000 (12:41 -0400)]
evas/main: shuffle shutdown order of filters and modules

Summary:
these both deallocate resources which can be needed during ecore_shutdown

ref df652673febd8617d3f458a5c14478534e927940
fix T7052

Reviewers: ManMower, devilhorns

Reviewed By: devilhorns

Subscribers: cedric, #committers

Tags: #efl

Maniphest Tasks: T7052

Differential Revision: https://phab.enlightenment.org/D6474

6 years agoeio/fallback: avoid emitting monitor events if the backend is pending deletion
Mike Blumenkrantz [Wed, 27 Jun 2018 12:48:59 +0000 (08:48 -0400)]
eio/fallback: avoid emitting monitor events if the backend is pending deletion

emitting events if the delete_me flag is set may result in events being emitted
for an already-freed monitor, resulting in both invalid memory access and a
deadlock later on if eio_shutdown has been called at this point

this causes the monitoring thread to check the status of the backend during the
block where the main loop and thread are in sync, avoiding any data races which
could occur when checking the flag at another time, and also avoiding accessing
the internals of the Ecore_Thread which could also have been deallocated during
shutdown

fix T7086

Differential Revision: https://phab.enlightenment.org/D6449

6 years agotests: check eio monitors for fallback monitoring
Mike Blumenkrantz [Tue, 26 Jun 2018 22:22:11 +0000 (18:22 -0400)]
tests: check eio monitors for fallback monitoring

the corresponding tests cannot be run when using fallback monitoring,
so be sure to skip them when fallback is detected to avoid erroneous
failure reporting

fix T7042

Differential Revision: https://phab.enlightenment.org/D6448

6 years agoeio: add method for determining if a monitor is using the fallback mechanism
Mike Blumenkrantz [Tue, 26 Jun 2018 22:20:37 +0000 (18:20 -0400)]
eio: add method for determining if a monitor is using the fallback mechanism

the fallback method of calling stat() on the monitored paths does not allow
for various eio events to be emitted, meaning that any application which relies
on those events can never receive them

this provides a method for checking a monitor to determine which functionality
is available, and also provides more explicit documentation regarding events
that are not provided by fallback monitoring

this method is marked as beta

@feature

Differential Revision: https://phab.enlightenment.org/D6447

6 years agotests: cancel the timeout timer when cleaning up eio tests
Mike Blumenkrantz [Tue, 26 Jun 2018 22:17:24 +0000 (18:17 -0400)]
tests: cancel the timeout timer when cleaning up eio tests

this timer could persist and cause cascading failures for subsequent
tests when running in non-forked mode

@fix

Differential Revision: https://phab.enlightenment.org/D6446

6 years agotests: increase eio monitor+sentry test initial delay to 0.05s
Mike Blumenkrantz [Tue, 26 Jun 2018 22:15:53 +0000 (18:15 -0400)]
tests: increase eio monitor+sentry test initial delay to 0.05s

this helps ensure that the fallback monitor can perform an initial
scan during the test while under load

fix T7042

Differential Revision: https://phab.enlightenment.org/D6445