To reproduce:
* elm_test -to "GenGrid Focus"
* turn on the "Focus Highlight" checkbox
* double-click an item to show a popup that should have the focus
As you can see the mouse-up event on the gengrid item steal the focus from the popup.
When, instead, the item is activated with the keyboard the popup correctly keep the focus.
I have this problem in both genlist and gengrid, so probably the error is not in the gengrid code, but somewhere else that I cannot find.
...any help appreciated
Summary:
Initially combobox used hover's default style, this patch would make combobox use hoversel_vertical style.
@Fix T3075 for 2nd point.
Signed-off-by: divyesh purohit <div.purohit@samsung.com>
Test Plan: please run combobox example from elementary_test
Reviewers: shilpasingh, cedric, raster
Subscribers: rajeshps, govi
Maniphest Tasks: T3075
Differential Revision: https://phab.enlightenment.org/D3596
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Before this patch, the filter was set inside the combobox. This is confusing for the
user of the object as it can't define a custom filter except on strings. It is also
impossible to bind as binding require to know what was put in and they can't differentiate
between a genlist filter being set from the binding to one being set internally by
the combobox.
@fix
genlist item is flaoting upon the expanded item when homoegeneous mode
is enabled.
when item block is recalculated by tree expanding, new item didn't
update their properties and values becuase of size hash, but edc
actaully need to be recaculated after swallowing tree size padding.
solve the issue by adding expand depth check in size hash.
you can check the previous issue in T2708
elementary_test->genlist tree->homoegeneous check
@fix
This @fix three (yes 3!) errors from commit cafda5928f9bf8116564219dc03d12c3a6e131b3
1. the initial value in the standard config is wrong, causing ctx menu to be disabled for everyone!
2. the argument of the api should be named "disabled" (like all the other disable API in this context)
3. (minor) fixed doc for elm_config_context_menu_disabled_set
NOTE: there are no configuration gui for this option, if you cleared your elm config recently you probably now have the ctx menu disabled everywhere, only solution is to wipe your config (to let elm create a new correct one)
Big SPANK to the author, the commiter and all the reviewers of D2463!!!
Summary:
While applying scroller on hoversel, several lines for item size
calculation were added, but they failed to have item get proper size.
Now redundant codes are removed and some are fixed, and hoversel will
be shown more well-ordered.
Test Plan: elementary_test hoversel
Reviewers: jpeg, raster, cedric
Differential Revision: https://phab.enlightenment.org/D3589
Summary:
This patch will reduce duplicate codes and code depth for readability
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Reviewers: seoz, Hermet, cedric
Subscribers: seoz
Differential Revision: https://phab.enlightenment.org/D3574
When using this function with wayland windows, we should be returning
the matching Ecore_Wl2_Window id, not the surface_id
ref T3058
@fix
Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
There is no need to recalc row and col position of the item, they are already stored in the item struct.
The old implementation was calculating wrong values and also was storing this wrong values in it->x and it->y, resulting in wrong results for the elm_gengrid_item_pos_get() function.
so, at the end, this is a @fix for the pos_get() function.
Also changed a bit the test to let the gengrid fill the window and thus testing behaviour on col/row changes.
I didn't find any regression in all the gengrid tests after this, let me know if it broke something for you.
Issue:
1. elementary_test -> slider
2. move mousewheel on slider and popup gets shown.
3. close the slider window very fast or increase the wheel_timeout to 4 seconds
4. observe crash
deleting subobjects here was occurring and triggering sizing eval
prior to removing the objects from the evas box, resulting in a failure
to accurately calculate and set size hints
also set delete_me to block sizing evals during subobj deletion loop
in order to avoid object thrashing during the time when nothing is happening
@fix
Summary:
When continous mouse clicks happen hover emits the hide signals continously
to edje. As the embryo program for hiding requires some time to finish the
animation, if within that time another mouse click comes the program gets
invoked again and thus the hiding will get delayed till the mouse clicks ends.
This patch fix that issue by adding a flag for emitting signals.
@fix T3006
Signed-off-by: godly.talias <godly.talias@samsung.com>
Test Plan: elementary_test
Reviewers: cedric, conr2d, prince.dubey, shilpasingh, Princekrdubey
Reviewed By: shilpasingh, Princekrdubey
Subscribers: Princekrdubey, divyesh, govi, rajeshps
Maniphest Tasks: T3006
Differential Revision: https://phab.enlightenment.org/D3526
Summary:
1) Recalc error occurs if combobox_hover_begin() was called by the app, it can be reproduced by keeping elm_combobox_hover_begin() in the test_combobox.c
2) Content_set in hover always considered "bottom" as it's best_location as it done during the constructor.
@fix
Signed-off-by: divyesh purohit <div.purohit@samsung.com>
Test Plan:
It can be reproduced by keeping elm_combobox_hover_begin() in the test_combobox.c.
Also the changing the position of combobox in the window would give top/bottom as it's best location.
Reviewers: cedric, shilpasingh
Reviewed By: shilpasingh
Subscribers: rajeshps, govi
Differential Revision: https://phab.enlightenment.org/D3544
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
it's wierd to expand hoversel without items.
the behavior made me suprised before i analyzed it.
it looked almost a bug.
if you have any other objections, please ping me.
conceptually, ctxpopup won't give any geometry information of the container.
this event is a way for returning the information of the container for user scenarios.
@feature.
fix @T2042
Coverity CID1347398 reports that this could potentially be a
non-portable assumption resulting in a sizeof mismatch. Fix this by
using the actual structure type in the call to calloc.
Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
This patch just fixes a compiler warning about using fw/fh
uninitialized if the conditional is false
Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
Davide Andreoli
divyesh purohit
Cedric BAIL
SangHyeon Lee
Jean-Philippe Andre
Jee-Yong Um
Minkyu Kang
Hermet Park
Chris Michael
Stefan Schmidt
Vincent Torri
Daniel Juyung Seo
Andy Williams
Amitesh Singh
Mike Blumenkrantz
godly.talias
Tom Hacohen
Minwoo, Lee
Carsten Haitzler
Larry