too unstabilized... need to fix up broken stuff...

@vtorri thanks for patching. but it occurs some troublesome both meson / autotools method... We can't apply this until one of them is stabilized as it does before.

it's inefficient to add the duplicated constant strings for the very rare case.
ERR("cannot allocate a digest context");

Please check a comment.

seriously...

looks correct it designed to print node tree info.

In D12194#233184, @jsuya wrote:

I have a question. In the case of _transit_effect_rotation_op, is anti_alias not necessary?

Please check my comments.

updated initialization variables.

Please check a comment.

looks fine to me, if there is any proper result of this, it would be nice.

Please check a comment.

Please check a comment.

Please check a comment.

This is not a common feature.

how this patch helps to this scenario:

good to me.

Please keep code convention. We prefer simpler way.

This looks good to me if agrresive optimization helps for better performance.

this looks wrong to me.
if user need the actual bound of the shape trees from the container, how they make it?