Summary:
The efl_gfx_color interface was not applied properly.
The implementation code of evas_object_smart_color_set was moved
to efl_gfx_color_set implementation code. But, these two functions are not same.
In efl_gfx_color_set impl, it has to call super's color set to apply
the given color values to widget object itself.
This bug caused color_set/get test failure and the following bug.
1. elm_image_add
2. evas_object_color_set
3. elm_image_file_set
4. show. See the given color is not applied. It was applied in the past.
Test Plan: color_set/get to elm_image object is included in test suite.
Reviewers: raster, cedric, herdsman, woohyun
Reviewed By: cedric
Subscribers: #committers, zmike
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D6163
Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
cl->socket is a Efl.Net.Socket_Simple. it is not inherit the Efl.Loop.fd class.
and The target of the class check have to be the inner_socket.
But inner_socket is Efl.Net.Socket_Ssl. Efl.Net.Socket_Ssl class is not inherit FD class.
Efl.Net.Socket_Tcp is inherit Efl.Loop.fd class. So, Need to add Efl.Net.Socket_Tcp to inheritance.
(The server side is a similar hierarchy. (ssl -> tcp -> ip -> fd))
Test Plan: N/A
Reviewers: cedric, jypark, myoungwoon, zmike, barbieri
Reviewed By: cedric
Subscribers: #committers, bowonryu, zmike
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D6168
Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
1. Scroller in scroller case, callback process should continue.
If child scroller does not scroll, parent scroller should be checked whether scroll or not.
So returning EINA_TRUE is necessary.
2. down.dragged flag is always EINA_FALSE, so scroll is never unlocked.
@fix
Test Plan:
precondition
elementary_config -> scrolling -> Set Scroll Smooth Amount to 0
elementary_test -> pagescroller
Reviewers: woohyun, zmike, singh.amitesh, cedric
Reviewed By: cedric
Subscribers: #committers, cedric, zmike
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D6187
Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
It was missing setting the 'is_ref' parameter to 1 to mirror the
internal_start call.
Test Plan: make check
Reviewers: cedric
Subscribers: #committers, zmike
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D6193
Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
Also changed the previously failing tests to use ck_assert_int_eqinstead
of fail_if for better reporting.
Test Plan: make check
Reviewers: cedric, felipealmeida
Reviewed By: cedric
Subscribers: #committers, zmike
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D6194
Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
There was wrong logic for calculating # of slices, dt, dist of each segment.
It caused bad rendering quality by putting too much slices on small text.
In addition, textpath didn't care about smoothness of curve's slope changing.
The patch fixes to check differences of previous points and next points for Evas Map.
So, textpath can show more smoothly curved text.
Also, it fixes "autofit" bug when text is much huge than given circle's size.
@tix
Test Plan:
- I'll attach screenshots for comparing rendering quality.
- To see "autofit" bug.
1. Run the following command.
ELM_ACCEL=gl ELM_SCALE=2.0 elementary_test -to "efl.ui.textpath"
2. Toggle "autofit" check box.
3. See the bug case. Textpath can't show all text properly.
Reviewers: raster, cedric, thiepha
Subscribers: #committers, zmike
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D6195
Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
Elm.List.Item lifecycle where an exception in Efl. They were trying to prevent the death of
there parent, to avoid dealing with safely walking on items list. This has been on the todo
list for years and is now fixed by this patch.
Ecore_Event can still be in flight while the client get freed by the server.
We are protected agaist that and keep the client partially alive until the event
are processed, but we need to disconnect from the server as soon as possible
as we might outlive the server.
Making wref disapear before the last event require additional work to use
them during the FREE event. As I think we should introduce a wref that
disapear after the invalidate state, it makes sense to move that to a later
stage. Especially because their is no specific event related to them
being set to NULL.
This is clearly a workaround to a bad situation, but there is no case
that I can find that solely lead to object being NULL nor is there any
code that should do that, but still it does in some case...
Technically I do not thing it is a correct behavior to force destroy
reference that evas do not hold, but evas_object are deaply tied to
the canvas they are build on and even after invalidate it is hard
to not have function call that would lead to crash. Making the pointer
incorrect thanks to eo indirection seems safer here.