I have this patch and the previous one properly updated, but I can't get phab to update it without failing in the middle. Once the first patches of the series are accepted, I can land all of it.
Showstopper Issues (15)
Rebase and rename.
Rebase and rename.
If a test needs to call ecore_shutdown then it should also call ecore_init after to restore the global state.
Rejected for being wrong and also for not having a commit log.
Rebase and fixup iterator.
Last questions. Do you have Focus split under the mouse?
In that case, is the mouse cursor over the new split or not?
and which version of EFL are you using?
How do you split? with the key bindings or through the UI?
Which parent are we talking about ?
I would be really carefull with item. things in here are *not* items.
This seems good, once the changes are submitted as a patch with authorship info then it can be merged.
I like that, but how about item instead of child? In the docs we talk about items inside the containers.
So... interface Efl.Ui.Arrangement with methods item_align and item_padding.
This would generate efl_ui_arrangement_item_align_set() and Efl.Ui.Table.ItemAlign.
Still needing a more in-depth analysis about the inpact of this patch in 'efl_part'.
Still working in some optimization aspects.
Avoid calling OWNERSHIP_UNIQUE while invalidating
Add a boolean flag for performance optmization reasons
I am wondering a bit about this revision. This is a legacy API, shouldn't something like this be used in the new scroller class ? Otherwise this feature will be gone and or unable to be used on new widgets.
Additionally, the definition of the actual header API is missing.
For me this could also imply that those properties do change the way the widget, where you set them on, gets placed. The attribute "child" is missing somehow ...
Can you adjust the commit title here to say something like: