I think it's complicating things at this point to deal with this. Unfortunately, it's not proper dateline wrapping as is (only arbitrary tile range extent wrapping).
With the frameState stuff the map renderer need not listen to view changes. The map renderer receives orders from the map. These orders are renderFrame_ calls with frameState objects representing the current state.
What we want in the end is vector tiles repeated just as raster
tiles. This change only avoids repeated tiles with the same
content being rendered and stored in the cache.
This needs more work still - see inline TODOs. It also has a
major bug - rendered vector features do not scale and do not
change their offset during panning. So only the initial view is
correct.
There is a ton of room for optimization here. The vector layer renderer should only render dirty areas and could maintain a cache of rendered canvas tiles. The vector source could have a simple spatial index for features (by tile coord). Need to also discuss how to work with this animation framework (to avoid the excess work while waiting for tiles to load on every other layer).
Currently only renders points as circles. Eventually will have methods for other geometry types and other symbolizer types. Also provides static methods to render point symbolizers (returning a canvas that can be drawn on another).
Tim Schaub
Éric Lemoine
ahocevar
Tom Payne
Frederic Junod