Instead of creating one GPtrArray per GtkSnapshotState and saving nodes
in there, create one GPtrArray per snapshot and assign a
start_node_index to every GtkSnapshotState as well as a n_nodes variable
so every state knows which nodes belong to it.
... and use it.
The function is a bit awkward because it requires 2 calls to
gtk_snapshot_pop(), but once you accept that, it's very convenient to
use, as can be seen by the 2 implementations.
We already take ints when setting the translation, so it can't
currently take any other values. Additionally, I was seeing large
costs in int -> double -> int for the rects in
gtk_snapshot_clips_rect(), as all callers really are ints (widget
allocations) and the clip region is int-based.
This change completely cleared a 2% rectangle_init_from_graphene from
the profile and is likely to have nice performance effects elsewhere
too.
This means we allocate the collect data with the state, avoiding
an extra allocation. Also, a union means every state object
is the same size and we could reuse the state objects.
It is now possible to call push() subfunctions for simple container
nodes with just a single child. So you can for example
gtk_snapshot_push_clip() a clip region that all the nodes that get
appended later will then obey.
gtk_snapshot_pop() will then not return a container node, but a clip
node containing the container node (and similar for the transform
example).
This is implemented internally by providing a "collect function" when
pushing that is called when popping to collects all the accumulated
nodes and combine them into the single node that gets returned.
To simplify things even more, gtk_snapshot_pop_and_append() has been
added, which pops the currently pushed node and appends it to the
parent.
The icon rendering code has been converted to this approach.
Instead of appending a container node and adding the nodes to it as they
come in, we now collect the nodes until gtk_snapshot_pop() is called and
then hand them out in a container node.
The caller of gtk_snapshot_push() is then responsible for doing whatever
he wants with the created node.
Another addigion is the keep_coordinates flag to gtk_snapshot_push()
which allows callers to keep the current offset and clip region or
discard it. Discarding is useful when doing transforms, keeping it is
useful when inserting effect nodes (like the ones I'm about to add).
Instead of having a setter for the transform, have a GskTransformNode.
Most of the oprations that GTK does do not require a transform, so it
doesn't make sense to have it as a primary attribute.
Also, changing the transform requires updating the uniforms of the GL
renderer, so we're happy if we can avoid that.
And use this to cull widgets and gadgets that are completely outside the
clip region.
A potential optimization is to apply this clip region to cairo contexts
created with gtk_snapshot_append_cairo_node(), but for that we'd need to
apply the inverse matrix to the clip region, and that causes rounding
errors.
Plus, I hope that cairo drawing becomes exceedingly rare so it won't be
used for the whole widget factory like today (which might also explain
why no culling happens in the widget factory outside the header bar.
We now try to emulate cairo_t:
We keep a stack of nodes via push/pop and a transform matrix.
So whenever a new node is added to the snapshot, we transform it
by the current transform matrix and append it to the current node.
