Which approach is better among D9535, D9556 and this one?

@cedric
(1) If it is possilbe and you do not mind, would you please let me know more about png_set_tRNS_to_alpha. I have no idea why the issue occurs without png_set_tRNS_to_alpha.
(2) The evas_image_load_file_data_png had used EINA_FILE_SEQUENTIAL for eina_file_map_all before the 9-patch commit. It could be another missing one. From the following change, your intention was that EINA_FILE_SEQUENTIAL should be used for pixels reading. Please make sure regarding this as well.

Use changed_map.
Remove incorrect(unnecessary) move, resize callback call

In D9476#177217, @Hermet wrote:

if other properties has been changed such as texture uv, image, point color, smooth option, antialiasing, etc
Does here it needs to check out all the property diff?

In D9476#177055, @Hermet wrote:

I guess this function won't be triggered if user just make it object visible again without any map update,

Does this work properly?

Fix crash issue and rendering issue on GL engine.
Remove incorrect error message. return value of ENFN->image_draw does not mean
that there is an unexpected error.

In D9353#173579, @Hermet wrote:

Is there any reason that filter only targets to group object? but not a single canvas object?
We can generally support filter for all types of graphical objects.

Set context clip using clipper.

do not use changed_map but add a condition to check spans->pts.

@Hermet
As we discussed at a meeting. I have checked if render_pre is called or not when rect's visibility is changed to hide from show.
The result is that the render_pre is called.
The procedure is following.

(1) hide rect
(2) _phase1_object_process, the is_active is true. 
   Please have a look into evas_object_is_active in this case. One of is_visible or was_visible is ture, then is_active could be true. 
   In case of smart object, there is a condition checking was_visible in _phase1_object_mapped.
(3) _phase1_object_changed_normal, push rect to render_objects
(4) _phase1_direct calls render_pre

Fix invalid memory read issue when obj size is changed.

Apply changes of children

Use offset correctly when filter surface draws to default surface

Why don't you add some more information such as "deprecated since 1.23 use elm_textpath_circular_set" ?

Use EINA_UNUSED correctly

same case on Efl.Ui.Slider

One more uninitialized variable from progressbar.

Fix one more uninitialized variable

elm_slider: fix uninitialized variable

Reviewed on https://review.tizen.org/gerrit/#/c/platform/upstream/efl/+/209046/ while phab was down.

Push restacked mapped object to restack_objs array

Enhance commit message

Use correct comment

The point here is, logically, the outbuf should be used as outter, not temporary. That's why we are saying it outbuf.

Now it's turned out that, curve requires temporary buffer always, grow buffer shouldn't use outbufer as temporary since it must have a temporary buffer.

Honestly I do not get what "Currently growbuf == outbuf is out of date." is. Do you mean that the outbuf is out of date, because it is same with growbuf? And is that for the first case that the outbuf is not dirty?

Update comment.

In D9085#166912, @Hermet wrote:

And It's very suspicious at this condition.

1288 if ((inbuf != outbuf) && out->dirty)

This condition might be related with this issue... Something make it out of the condition ...

Use a temp buffer on grow_add side as well to reduce calling blend_add based on
valuable hermet's comment.

Let me know more about 'map object'. Does it mean the object using Evas_Map such as button of evas_object_map_set(button, map)?

Use correct name blendcmd instead of curvecmd

It would be better if we have enhanced documentation and related TC as well. Thank you.

Enhanced based on comment. This is much more neat!

Update based on comment.

Update based on comment.

Remove unintended lines.

Update based on comments.

In D8892#162737, @Hermet wrote:

1. I've considered for the interface, how about using align_hint ?

If the "using align_hint" means kind of "using Efl.Gfx.Hint.hint_align", it could be more flexible than using these enums. But it seems that you did not mean Efl.Gfx.Hint.hint_align because this is hint on how to align an object inside its container.

Remove unrelated changes

Update based on a comment saying make it simpler

fix typo

+1024

The warning probably comes from

@bu5hm4n
(1) Are you using gl engine?
(2) What is the result of eina_btlog?
(3) The result would be not caused by incorrect example.

(Oooops... sorry...) Warnings are removed.

sorry, this is a change to test arc

Remove unnecessary changes

Amend based on comments

Um... no details?

Amend based on comments

Add this to the meson build file