Derek Foreman [Tue, 13 Feb 2018 23:09:19 +0000 (17:09 -0600)]
ecore_wl2: Fix memory leak on shutdown
The output list wasn't properly cleaned up on shutdown/disconnect.
Mike Blumenkrantz [Tue, 13 Feb 2018 22:28:40 +0000 (17:28 -0500)]
efl_selection_manager: shut up shut up shut up ecore-x when running in wayland
fix T6643
Mike Blumenkrantz [Tue, 13 Feb 2018 22:23:44 +0000 (17:23 -0500)]
efl_gesture_manager: free hash in destructor
every canvas object has had this leak for some time:
==25217== 64 bytes in 1 blocks are definitely lost in loss record 1,669 of 3,515
==25217== at 0x4C2FB6B: malloc (vg_replace_malloc.c:299)
==25217== by 0x688D839: eina_hash_new (eina_hash.c:755)
==25217== by 0x688D940: eina_hash_string_superfast_new (eina_hash.c:791)
==25217== by 0x6265F8C: _efl_gesture_manager_efl_object_constructor (efl_gesture_manager.c:66)
Daniel Kolesa [Tue, 13 Feb 2018 20:48:46 +0000 (21:48 +0100)]
eolian: properly initialize warned (to avoid random parse failures)
Romain Naour [Mon, 5 Feb 2018 14:07:15 +0000 (15:07 +0100)]
elua: add eina in environment variable library path
In-tree libeina.so library path is missing while running elua.
http://autobuild.buildroot.net/results/2bb/
2bbe89694c206d764d51a5dd9c75bc4aa844e98d/build-end.log
@ref T6670
Cc: Daniel Kolesa <d.kolesa@samsung.com>
Signed-off-by: Daniel Kolesa <d.kolesa@osg.samsung.com>
Chris Michael [Tue, 6 Feb 2018 15:58:58 +0000 (10:58 -0500)]
evas drm engines: Fix variable type for output
We already include the Ecore_Drm2 header for these engines, so there
is no need for the 'output' field to be a void pointer here.
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Chris Michael [Tue, 6 Feb 2018 14:12:39 +0000 (09:12 -0500)]
ecore-evas-drm: Use output viewport in ecore_evas_animator_tick calls
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Chris Michael [Tue, 6 Feb 2018 14:07:26 +0000 (09:07 -0500)]
ecore-drm2: Add API functions to get/set output relative mode
These API functions will be used inside Enlightenment in order to
determine the proper extended screen size based on outputs relative
position.
@feature
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Chris Michael [Tue, 6 Feb 2018 13:49:46 +0000 (08:49 -0500)]
ecore-evas-drm: Include Evas_Internal header
We need to include this header so that we can use the new
efl_canvas_output functions
Signed-off-by: Chris Michael <cp.michael@samsung.com>
WooHyun Jung [Tue, 13 Feb 2018 13:50:14 +0000 (22:50 +0900)]
efl_ui_nstate: make "changed" callback is called properly
- "changed" callback is not called when theme_apply is called
- elm_check_state_set does not call "changed" callback to keep
backward compatibility
WooHyun Jung [Tue, 13 Feb 2018 10:20:19 +0000 (19:20 +0900)]
Revert "elm: do not call "changed" callback when setting style"
This reverts commit
b1ee35e35ce3132566244abbbd3a804167ef9c00.
WooHyun Jung [Tue, 13 Feb 2018 09:39:20 +0000 (18:39 +0900)]
Revert "Efl.Ui.Nstate: do not call "changed" callback"
This reverts commit
fd8e9f9c4e8dbda6c379dc464f562f5b2c681d31.
Daniel Kolesa [Tue, 13 Feb 2018 09:14:49 +0000 (10:14 +0100)]
eolian: print all duplicate warnings before aborting build
Instead of aborting at the first warning, accumulate the warnings
and stop at the end of validation.
Carsten Haitzler (Rasterman) [Tue, 13 Feb 2018 09:12:11 +0000 (18:12 +0900)]
elm focus ... set focus in focus in if focus is not already on something
let's actually set focus in focus_in on window (ecore evas) IF focus
is not set. this seems to make terminology still work...
Carsten Haitzler (Rasterman) [Tue, 13 Feb 2018 06:28:47 +0000 (15:28 +0900)]
efl ui win - if job runs after destruction of evas so dont flush pending
Carsten Haitzler (Rasterman) [Tue, 13 Feb 2018 06:21:49 +0000 (15:21 +0900)]
elm - dont get top widget for non-ui objects
Carsten Haitzler (Rasterman) [Tue, 13 Feb 2018 04:46:38 +0000 (13:46 +0900)]
efl ui win - fix focus break for terminology (at least)
at least terminology was broken by this. perhaps other efl apps. if an
efl app has decided to manually play with focus itself, like set focus
to some obj then leav ie, this line breaks that and interferes with
the object focused whenever the window gets focus set. if there is a
focus manager at least this is a new thing and so let the focus
manager set focus, but dont set old evas focus.
this fixes the break added in
56beb861e83cbc830c82db7473e0a88461e2e213
Mike Blumenkrantz [Mon, 12 Feb 2018 22:14:34 +0000 (17:14 -0500)]
evas_vg: implement interpolate from Efl.Gfx.Path
Mike Blumenkrantz [Mon, 12 Feb 2018 21:58:57 +0000 (16:58 -0500)]
interfaces: merge all bounds_get methods into efl.gfx.path
ref
884fc35160152712decdf7fd43082ef3912c0856
Mike Blumenkrantz [Mon, 12 Feb 2018 19:29:20 +0000 (14:29 -0500)]
evas3d: use a type setter function with finalize instead of a custom constructor
resolves a conflict with the 'constructor' method name
Mike Blumenkrantz [Mon, 12 Feb 2018 19:28:45 +0000 (14:28 -0500)]
evas3d: add "unknown" node type for use with errors
Mike Blumenkrantz [Mon, 12 Feb 2018 21:02:06 +0000 (16:02 -0500)]
interfaces: remove Efl.Text.Properties
this is provided in Efl.Text.Font
Mike Blumenkrantz [Mon, 12 Feb 2018 20:34:01 +0000 (15:34 -0500)]
efl_text_font: copy docs from efl_text_properties
Mike Blumenkrantz [Mon, 12 Feb 2018 20:04:15 +0000 (15:04 -0500)]
efl_vg: use efl.object's name property instead of implementing a new one
Mike Blumenkrantz [Mon, 12 Feb 2018 18:45:07 +0000 (13:45 -0500)]
ector: rename ector_renderer "bounds_get" method to "boundary_get"
this conflicted with a method from Efl.Gfx.Path
Mike Blumenkrantz [Mon, 12 Feb 2018 18:23:25 +0000 (13:23 -0500)]
ector: fix property/method name conflict for "fill"
Shape implements a property, sw/gl/cairo implements a method
rename method to op_fill
Mike Blumenkrantz [Mon, 12 Feb 2018 17:56:31 +0000 (12:56 -0500)]
eolian: convert all of cedric's tabs to spaces in .eo files
Daniel Kolesa [Mon, 12 Feb 2018 17:56:38 +0000 (18:56 +0100)]
eolian: stop build with EOLIAN_WARN_FUNC_DUPLICATES > 1
Derek Foreman [Mon, 12 Feb 2018 17:27:11 +0000 (11:27 -0600)]
ecore_wl2: Send window hide in response to xdg popup done
Clients should be sent this event when the compositor sends popup done.
Marcel Hollerbach [Sun, 11 Feb 2018 18:04:06 +0000 (18:04 +0000)]
efl_ui_win: move inital focus set to focus_in handler
it seems that focus changes to FOCUS=FALSE are causing autodel windows
to kill themself, so we only set the focus on the window if the window
manager calls focus in AND we dont have anything to focus and nothing is
focused yet.
Carsten Haitzler (Rasterman) [Sun, 11 Feb 2018 10:20:46 +0000 (19:20 +0900)]
edje - entry - fix empty item handling
tjhis should fix empty items which might have a null item string...
deal with it
fixes T6668
@fix
Carsten Haitzler (Rasterman) [Sun, 11 Feb 2018 06:55:13 +0000 (15:55 +0900)]
eina - tiler -= return true if rect alreadt there as its not a failure
docs say return true on succesas, false on failure. adding a rect we
already added is not a failur. it's an optimization to a NOP. so fix.
this was brought up by and fixes T6669 ... but in the opposite way.
Carsten Haitzler (Rasterman) [Sun, 11 Feb 2018 05:53:28 +0000 (14:53 +0900)]
elm - fix new em_config func prototypes in header to have real param decl
Carsten Haitzler (Rasterman) [Sat, 10 Feb 2018 07:26:15 +0000 (16:26 +0900)]
efl ui format - fix format checking to be correct
the format string checking was just ... wrong. it didnt' handle %%. it
didn't handle constant strings with no format in them... now it does.
fixes T6697
@fix
Carsten Haitzler (Rasterman) [Sat, 10 Feb 2018 07:24:11 +0000 (16:24 +0900)]
elm progressbar - fix val set
set initial min and max to 0 and 1 respectively so val set works at
the start...
fixes T6697
@fix
Carsten Haitzler (Rasterman) [Sat, 10 Feb 2018 05:38:29 +0000 (14:38 +0900)]
elm - make progress test code a little more solid and nice to read
use sizeof the buffer rather than constant in snprint and 0.5 rather
than .50 which almost looks like a typo... :) also use correct escape
for % (%%%% so it becoems %% which is correct for a fmt string yto
produce %)
Derek Foreman [Fri, 9 Feb 2018 17:25:41 +0000 (11:25 -0600)]
elm: Fix double shutdowns of ecore_x and ecore_wl2
This is now done elsewhere. wayland clients were ERRing on shutdown.
ref
9cf61c2f6d215e3adfa1567154a229eb49116769
Derek Foreman [Fri, 9 Feb 2018 17:11:24 +0000 (11:11 -0600)]
wayland_imf: Clear hide timer when deleting context
Otherwise the callback can fire after the module is unloaded leading to some
very entertaining to debug crashes at shutdown.
Stephen Houston [Fri, 9 Feb 2018 16:21:55 +0000 (10:21 -0600)]
Elm Progressbar test: Add in examples of efl regressions to progressbar.
Derek Foreman [Fri, 9 Feb 2018 15:49:52 +0000 (09:49 -0600)]
ecore_wl2: Destroy offscreen event handler in surface destroy
Or we can get the event after destroying the surface and crash.
Lauro Moura [Thu, 1 Feb 2018 18:38:48 +0000 (15:38 -0300)]
efl_mono: Distcheck fixes
- Export required sources
- Avoid generated sources being passed as static ones
Lauro Moura [Tue, 6 Feb 2018 19:10:36 +0000 (16:10 -0300)]
eina_mono: Avoid calling eina_init directly from the modules.
They must be initialized from the eina.Config.Init() method.
Stefan Schmidt [Thu, 8 Feb 2018 09:45:24 +0000 (10:45 +0100)]
ci: make sure we fail the build and pass on the non zero exit code
So far we just kept going on when one of the make targets failed with
the next one.
Sungtaek Hong [Thu, 8 Feb 2018 09:37:51 +0000 (11:37 +0200)]
efl_part: call Efl.Text of each class' part
Summary:
Macro ELM_PART_TEXT_DEFAULT_IMPLEMENT calls
super class' Efl.Text interface, rather than this class.
Reviewers: jpeg, woohyun, Jaehyun_Cho, herdsman
Subscribers: cedric, id213sin
Differential Revision: https://phab.enlightenment.org/D5799
Sungtaek Hong [Thu, 8 Feb 2018 05:53:59 +0000 (14:53 +0900)]
efl_part: use efl_ui_widget's default_part_get
Summary:
efl_part macros are using each widget's internally defined
default_part_get() functions to get default part name.
This might potentially cause errors when future widgets
inherits the widget but not overriding Efl.Text.text and
Efl.Content.content.
Reviewers: jpeg, cedric, woohyun, Jaehyun_Cho
Differential Revision: https://phab.enlightenment.org/D5797
subhransu mohanty [Thu, 8 Feb 2018 05:27:15 +0000 (14:27 +0900)]
ector: Updated the ector_buffer_pixels_set() api with stride info
Reviewers: jypark, jpeg
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D5795
Mike Blumenkrantz [Wed, 7 Feb 2018 18:32:08 +0000 (13:32 -0500)]
elm: remove ecore-x init from quicklaunch post-fork
this is invalid for basically every possible case. allow components
to naturally call this function as needed
@fix
Mike Blumenkrantz [Wed, 7 Feb 2018 18:30:37 +0000 (13:30 -0500)]
elm: call ecore_app_args_set() after quicklaunch fork
ensure that apps get expected behavior when calling args_get()
@fix
Mike Blumenkrantz [Wed, 7 Feb 2018 17:09:35 +0000 (12:09 -0500)]
ecore-imf: do ecore-x init during ctx creation instead of module init
this avoids a number of issues with efl under wayland, msot notably the
automatic triggering of xwayland during every app startup
@fix
Mike Blumenkrantz [Wed, 7 Feb 2018 17:03:41 +0000 (12:03 -0500)]
elm: move ecore-x and ecore-wl2 connection init to selection manager
this seems to be the only place where the related components are
explicitly used
neither of these components have fork-safe connections, so there is no
benefit to calling them during quicklaunch init
Mike Blumenkrantz [Wed, 7 Feb 2018 17:03:00 +0000 (12:03 -0500)]
elm: print error if setsid() fails during quicklaunch fork
Mike Blumenkrantz [Wed, 7 Feb 2018 16:55:40 +0000 (11:55 -0500)]
eio: remove broken pid-based monitor reset copied from ecore-file
this has no technical basis for existing and should be handled by
individual monitor backends
@fix
Mike Blumenkrantz [Wed, 7 Feb 2018 16:50:57 +0000 (11:50 -0500)]
ecore: fix event flushing to not erase+leak events
stealing the message data here prevents events which aren't being flushed from
ever being usable again and is unnecessary since the free callback will be
automatically called during the destructor
ref
5dd52fd09b7d79c70b3134423a87aa6400a2d994
Vincent Torri [Wed, 7 Feb 2018 13:10:24 +0000 (22:10 +0900)]
Evas: remove depth usage in GDI and DirectDraw engines
Test Plan: compilation and elementary_test
Reviewers: raster
Reviewed By: raster
Subscribers: cedric, raster
Differential Revision: https://phab.enlightenment.org/D5792
Jaehyun Cho [Wed, 7 Feb 2018 11:51:04 +0000 (20:51 +0900)]
elm_part_helper: Fix to use registered default text part
Shinwoo Kim [Wed, 7 Feb 2018 07:18:53 +0000 (16:18 +0900)]
Efl.Ui.Nstate: do not call "changed" callback
Before solving following problem in Efl.Ui.Check
https://phab.enlightenment.org/T6673
The changed callback is called with opposite value,
if application is using Efl.Ui.Check with state pointer.
The reason is that the changed callback is called in
efl_ui_nstate_activate, and the value refered by state pointer
is changed only after the efl_ui_nstate_activate is called.
static void
_activate(Evas_Object *obj)
{
EFL_UI_CHECK_DATA_GET(obj, sd);
efl_ui_nstate_activate(obj);
if (sd->statep) *sd->statep = efl_ui_nstate_value_get(obj);
Jiyoun Park [Wed, 4 Apr 2018 04:43:12 +0000 (13:43 +0900)]
elm_genlist: fix rebase build error related with on_focus_update to Focus.Object
widget: Move on_focus_update to Focus.Object
Following @taxi2se's recommendation. This is indeed a focus method, and
Widget already inherits from Focus.Object.
Ping @bu5hm4n who probably wants to adapt this further.
Ref T5363
Jiyoun Park [Wed, 4 Apr 2018 01:41:22 +0000 (10:41 +0900)]
efl_spec: fix build error related with dmabuf of ecore_wl2
Jiyoun Park [Wed, 4 Apr 2018 01:39:05 +0000 (10:39 +0900)]
ecore_wl2_tbmsurface: modify tbmsurface based on ecore_wl2_surface change of opensource
Chris Michael [Tue, 6 Feb 2018 18:01:25 +0000 (13:01 -0500)]
ecore-drm2: Ensure output changed event is sent for removed outputs
If an output gets disconnected, we would still like to be notified of
that. In the _outputs_update function, we mark an output as
disconnected by setting output->connected and output->enabled to
FALSE. With this line in place (in the output_event_send function) we
would never get notified if an output was disconnected.
@fix
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Carsten Haitzler (Rasterman) [Tue, 6 Feb 2018 16:10:42 +0000 (01:10 +0900)]
disable async mode (use sync mode) for ibus when keymap changes
so scenario:
1. use ibus
2. have at least english input and japanese input (or korean etc.)
3. have 2 kbd layouts (english and greek).
4. enable "use system keyboard layout" in ibus advanced settings
5. switch to english input mode
6. switch to greek key layout
7. type and get english, not greek input as you should
@fix for both terminology and elm/efl entry/ytext input.
Shinwoo Kim [Tue, 6 Feb 2018 13:38:14 +0000 (22:38 +0900)]
elm: do not call "changed" callback when setting style
Amitesh Singh [Tue, 6 Feb 2018 13:33:58 +0000 (22:33 +0900)]
efl.ui.progressbar: add checks for min & max value
this improves the validity check for min & max values
provided from user side.
Jean-Philippe Andre [Tue, 6 Feb 2018 12:23:45 +0000 (21:23 +0900)]
efl_ui: Disable includes that don't compile
Efl.Ui.Text:
The EO file contains a lot of references to legacy Elm types, which are
defined in elm_general. They should be checked and moved over to
efl_ui.eot if necessary.
Efl.Ui.Multibuttonentry:
This class was originally supposed to be based on a Model Item but as of
now the API is still uncertain, so MBE itself hasn't been worked on
more. Disable this from EO-only apps until its API is fixed.
Ref T6666
Shinwoo Kim [Mon, 5 Feb 2018 11:02:06 +0000 (20:02 +0900)]
eina_tiler: fix typo of rect comparing
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
eio: make inotify monitors fork-safe
@fix
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
eeze: simplify watch code and make it fork-safe
avoid reads from parent process's udev fds
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
ecore-file: make monitoring truly fork-safe
add a fork reset callback and attempt to preserve existing monitors
during reset
@fix
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
elm: call ecore_event_init() during init
events from this component are used internally in elm, don't rely on
other components to do it
@fix
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
ecore: make dbus-using modules fork-safe
reset dbus connections to ensure continued functionality
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
eldbus: make connections fork-safe
after a fork, any existing connection objects can no longer be used,
but it's up to the user to destroy them. internally, this prevents
existing connections from ever being returned as valid connections
and creates new ones after a fork
also destroy fd handlers for connections to ensure that no data is
accidentally clobbered before the connections are cleaned up
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
efreet: reset ipc connection after fork
ensure that existing connection is not reused
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
eldbus: do not require error return when adding/removing signal matches
this forces the functions to be synchronous and blocks (potentially forever)
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
quicklaunch: use XDG_RUNTIME_DIR instead of hardcoded /tmp
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
efl-wl: add function for passing ecore-exe flags to launched exes
@feature
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
efl-wl: no need to check exe tag on app del, just remove pid always
Mike Blumenkrantz [Fri, 2 Feb 2018 21:59:56 +0000 (16:59 -0500)]
efl-wl: add functions for managing allowed client pids
@feature
Shinwoo Kim [Fri, 2 Feb 2018 10:07:04 +0000 (19:07 +0900)]
elm: fix typo - remove line to debug
Shinwoo Kim [Fri, 2 Feb 2018 09:56:58 +0000 (18:56 +0900)]
elm: use Dbus securely
The Efl.Access.Attribute is using key and value.
The value could be NULL. If the value is NULL, then following error occurs.
*error:
arguments to dbus_message_iter_append_basic() were incorrect,
assertion "_dbus_check_is_valid_utf8 (*string_p)" failed
in file ../../dbus/dbus-message.c line 2712.
This is normally a bug in some application using the D-Bus library.
Array or variant type requires that type string be written, but end_dict_entry
was written.
The overall signature expected here was 'a{ss}' and we are on byte 3 of that
signature.
Radoslaw Cybulski [Fri, 2 Feb 2018 07:53:42 +0000 (08:53 +0100)]
elm: fix for invalid attribute matching
Summary:
Fixes invalid attribute matching in at-spi's collection interface's GetMatches***
functions.
Reviewers: stanluk
Reviewed By: stanluk
Subscribers: lukasz.stanislawski, cedric
Differential Revision: https://phab.enlightenment.org/D5774
Vincent Torri [Thu, 1 Feb 2018 19:28:57 +0000 (20:28 +0100)]
evil: final step : remove completely the binary, useless now that the unit test is theree
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Carsten Haitzler (Rasterman) [Thu, 1 Feb 2018 11:09:05 +0000 (20:09 +0900)]
efl ui did - dont segv if createicon func is not provided
Shinwoo Kim [Thu, 1 Feb 2018 05:20:20 +0000 (14:20 +0900)]
Use ERR instead of CRI if *DATA_GET* returns NULL
This patch set is for remains.
Shinwoo Kim [Thu, 1 Feb 2018 03:16:28 +0000 (12:16 +0900)]
elm: Use ERR instead of CRI if *DATA_GET* returns NULL
Vincent Torri [Wed, 31 Jan 2018 12:35:55 +0000 (13:35 +0100)]
eina tests: add missing semicolumn
Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
Vincent Torri [Wed, 31 Jan 2018 12:35:03 +0000 (13:35 +0100)]
evil: remove useless binary tests
Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
Daniel Kolesa [Wed, 31 Jan 2018 18:14:15 +0000 (19:14 +0100)]
eolian: no need to store 'parsed' table anymore either
Daniel Kolesa [Wed, 31 Jan 2018 18:06:37 +0000 (19:06 +0100)]
eolian: simplify parsed checking
Daniel Kolesa [Wed, 31 Jan 2018 17:10:06 +0000 (18:10 +0100)]
eolian: no need to track currently parsing file anymore
Daniel Kolesa [Wed, 31 Jan 2018 17:06:17 +0000 (18:06 +0100)]
eolian: all dependency parsing is now deferred
YeongJong Lee [Wed, 31 Jan 2018 14:04:26 +0000 (15:04 +0100)]
efl_ui_focus_parent_provider_gen: code refactoring
Summary: remove needless function call and else statement
Reviewers: bu5hm4n
Subscribers: cedric, jpeg
Differential Revision: https://phab.enlightenment.org/D5732
YeongJong Lee [Wed, 31 Jan 2018 13:29:18 +0000 (14:29 +0100)]
efl_ui_focus_manager_sub: avoid a call to NULL
Summary:
we need to consider the manager change to NULL
@ref T6616
Reviewers: bu5hm4n
Subscribers: cedric
Maniphest Tasks: T6616
Differential Revision: https://phab.enlightenment.org/D5771
YeongJong Lee [Wed, 31 Jan 2018 13:29:12 +0000 (14:29 +0100)]
elm_menu: avoid a call to NULL
Summary: the provider_find can be called before parent_set in constructor.
Test Plan:
1. EINA_LOG_LEVELS=eo:2 elementary_test -to menu
2. check that there is no warning message about efl_provider_find
Reviewers: jpeg, bu5hm4n
Reviewed By: bu5hm4n
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D5773
Marcel Hollerbach [Sat, 27 Jan 2018 16:57:43 +0000 (17:57 +0100)]
efl_ui_focus_composition: remove elements_flush
this can also be done when calling prepare_logical
Amitesh Singh [Wed, 31 Jan 2018 09:04:39 +0000 (18:04 +0900)]
Efl.ui.progressbar: remove duplicate call
eina_value_set() was already getting called
Amitesh Singh [Wed, 17 Jan 2018 06:52:58 +0000 (15:52 +0900)]
Efl.Ui.Progressbar: implement range min max
Derek Foreman [Tue, 30 Jan 2018 23:06:48 +0000 (17:06 -0600)]
wl2_surface_dmabuf: Trim the buffer queue after a while
If we have more buffers than we need for 100 frames then drop the oldest.
This can happen if we're on a hardware plane and then removed from it, or
really whenever the compositor feels like holding onto a few frames.
Trimming the queue too soon could result in having to do a costly full
frame redraw, so we wait a while to make sure we don't need one again.
Having more frames than we need costs us a little every draw since we
always use the oldest available. It also wastes memory.
Derek Foreman [Tue, 30 Jan 2018 22:32:45 +0000 (16:32 -0600)]
ecore_wl2_surface: Flush surfaces for offscreen windows
When a surface leaves all outputs we can discard its buffers to save
memory.
Currently most compositors don't send leave events for iconify, so this
pretty much just saves us a cursor buffer under weston for now, but in
the future it could be used for freeing resources of offscreen (fully
occluded or iconified) windows.
Derek Foreman [Tue, 30 Jan 2018 22:29:32 +0000 (16:29 -0600)]
wl2_surface_dmabuf: improve surface flush
We used to abandon all buffers even if they were locked. This can't
actually free them until they're all released by the compositor.
Instead just free any buffer the compositor doesn't have locked, so we can
still re-use them when they're released without needing a full redraw.
There's probably room for additional cleverness here. If we get a new
frame event before the buffer release we may want to keep it, and if we
get the release first we may want to abandon it.
Derek Foreman [Tue, 30 Jan 2018 21:19:43 +0000 (15:19 -0600)]
ecore_wl2: Add new event for windows not displayed on any output
When a window leaves all outputs we can free its render buffers to save
memory. This new event is generated when a window leaves all outputs.