Summary:
1) According to description in .eo file,
some APIs return not 0 when the given object is NULL or not proper object.
But, several APIs are not implemented as those description.
Now, they will return a value as description.
2) If there is no description about the return value when the API fail,
put description according to pair API or old version of elementary.
@fix
Reviewers: woohyun, Hermet, seoz
Differential Revision: https://phab.enlightenment.org/D1933
Summary:
Resolved FIXME filled all smart callbacks and also set it in the constructor
Signed-off-by: kabeer khan <kabeer.khan@samsung.com>
Reviewers: seoz
Differential Revision: https://phab.enlightenment.org/D1876
the elm config tool only lets you set accel preference. if you set
none, then the old engine config takes charge, but there is no way to
change/set this, so deprecate the old engine set and don't use it.
this accel preference has been around for a while, so use it now
instead.
Summary: Added indicator_visible_mode to each slider in order not to make elm_config controlls every sliders.
Reviewers: Hermet, woohyun, seoz
Subscribers: seoz
Differential Revision: https://phab.enlightenment.org/D1887
Summary:
When ecore_evas_window_profile_supported_get API return EINA_FALSE,
elm_win widget should not get profile from ecore_evas_window_profile_get API.
@fix
Reviewers: woohyun, Hermet, seoz
Reviewed By: seoz
Differential Revision: https://phab.enlightenment.org/D1943
while this is technically an abi break, it should actually have no
real impact. it will have warnings - sure, but running existing code
will simply not see the extra parammeters on the stack and not use
them so it'll continue to work. if we removed paramms, we'd be in
pain. (same applies to params passed in registers).
this fixes T2031
Summary:
There is a bug that text disappears when theme is changed after label shown.
This patch fixed it.
@fix
Reviewers: woohyun, Hermet, Jaehyun, seoz
Subscribers: Jaehyun, seoz
Differential Revision: https://phab.enlightenment.org/D1923
Summary:
tooltip could be hide by the other object which have lower layer than target object of tooltip.
the layer of content could be changed, when is set to the tooltip as a content.
@fix
Test Plan: check added test lines
Reviewers: cedric, jpeg, seoz, raster
Reviewed By: raster
Subscribers: singh.amitesh, stefan_schmidt, seoz
Differential Revision: https://phab.enlightenment.org/D1889
Summary:
Label should not start sliding before elm_label_slide_go() is called.
However, label starts sliding automatically,
because resize callback emits slide start signal.
This patch prevents label from sliding before calling elm_label_slide_go().
@fix
Reviewers: woohyun, Hermet, seoz
Subscribers: CHAN, woohyun
Differential Revision: https://phab.enlightenment.org/D1906
Summary:
Eina_List from elm_gengrid_realized_items_get must be freed by caller,
but elm_gengrid_realized_items_update doesn't free Eina_List.
So memory leak is happens.
@fix
Test Plan: Call elm_gengrid_realized_items_update repeatly and check memory share increase.
Reviewers: raster, seoz
Differential Revision: https://phab.enlightenment.org/D1903
Summary:
It could success only when the widget was visible.
The API must return EINA_TRUE if buffer and size has no problem, not visiblity of widget.
@fix
Reviewers: Hermet, woohyun, jpeg
Differential Revision: https://phab.enlightenment.org/D1917
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
When layout clicked, "elm_widget_on_focus()" called with "entry_focus_in()" callback.
The "view update" and "focus set to entry" already call in those callbacks, so i deleted duplicate code and it also make focus issue when user click the entry to add item.
@fix
Test Plan:
Run elementary_test.
1. excute multibuttonentry.
2. click the entry area.
3. see the elm_widget_on_focus called twice.
(layout callback function called view update -> elm_box_unpack called in view update ->
entry lose focus -> on_focus_callback call again)
Reviewers: Hermet, seoz, Jaehyun_Cho, JackDanielZ
Differential Revision: https://phab.enlightenment.org/D1902
Summary: The API must return ELM_WIN_UNKNOWN (-1) when the given object is improper.
Test Plan: None.
Reviewers: woohyun, JackDanielZ, seoz
Differential Revision: https://phab.enlightenment.org/D1907
Use the min between w and h of the evas object to
request the size in the icon theme. Fdo icon are always
square (1:1) so min seems correct.
Also improved the test to include a test that do
not relay on the size_hint_min of the icon.
Now in all the test I have done always the correct
fdo icon are used.
This reverts commit 75176a789981bd6b1ef0bfcc95793c896b5bde07.
Was not the correct solution, as min_size can be smaller than
the real wanted size. Need to find another solution.
Thanks TAsn to spot this
This @fix fdo icons to load the correct image for the requested size.
Whitout this the small images (16x16) was never used, because once a big
(64x64 for example) image was loaded, that size was used as min.
See the Icon Standard test for a better explanation
The code there was trying to reload the icon to have an
image that should match the requested size, this was totally
wrong, the calculation was always resulting in an image bigger
than the needed.
It was also useless as this logic is yet handled in the resize
callback.
You can now select the size of the icons in the list and
also added a single icon that you can resize.
The size of the icons is important as icon themes usually
provide different images for different sizes.
This test now spot out some errors in elm_icon wrt icon
sizing, will fix soon
This fixes the side effect that introduced from 1e06309be42fd2b5d0900b16da0178215c111681
Since transit has supported adding a map effect on the previous map status,
transit tried to access object-state map pointer.
but theses map pointer won't be valid if the transit keeps the last state
because it doesn't back up any object-state data.