Introduce a second level of immutability, over each of the collections held by a style: sources, images, and layers. Tracking immutability at this level allows us to short-circuit significant portions of the RenderStyle update logic via a simple equality check, greatly improving performance.
* Eliminate updateBatch in favor of diffing layers and detecting changes to properties upon which layout depends.
* Replace RenderSource::{update,remove,invalidate,reload}Tiles with a single update method
* Replace TilePyramid::{update,remove,invalidate,reload}Tiles with a single update method
* Remove Style& dependency TODO from GeometryTile and TileParameters
* Eliminates the need for EnableImmutableFromThis
* Eliminates the dependency of {Source,Layer}::Impl on corresponding Render class (circular dependency)
* Don't use it to track zoom changes. Instead, Style::update can use the zoom history to check for a change in zoom from the previous frame.
* Don't use it to track active property transitions. Style already knows which layers/light have an active transition, and can re-evaluate only those that do.
This leaves layer property changes as the only use of RecalculateStyle.
Combine Style::cascade, recalculate, relayout, and updateTiles into a single method. This allows multiple loops over sources and layers to be consolidated and prepares for additional Style-Map decoupling: rather than tracking pending updates via a set of Update flags held by the Map and passed to the Style, the Style can use its own data to determine what to update.
Moves the following to the renderer directory and out of the style namespace:
* CascadeParameters
* PropertyEvaluationParameters
* UpdateParameters
* PropertyEvaluator
* DataDrivenPropertyEvaluator
* CrossFadedPropertyEvaluator
* PaintPropertyBinder
* PaintProperyStatistics
* PossiblyEvaluatedPropertyValue
* TransitioningLight
* EvaluatedLight
This reverses #3095. Rationale:
* We're now exposing source attributes as a public API. Making those attributes unavailable at certain times complicates that API.
* We're preparing to split RenderSource out of Source. Removing this removes a point of coupling between the two.
* Simplify SymbolLayout; it never needs to care about more than one SpriteAtlas.
* Move the reference from SymbolLayer::Impl to SymbolBucket. This is a prerequisite for making layer Impls immutable.
- Expose glyph and icon information to workers via message interface.
- Glyph/SpriteAtlas track which tiles have outstanding requests
and send messages to them when glyphs/icons become available.
- Remove obsolete "updateSymbolDependentTiles" pathway
- Symbol preparation for a tile now depends on all glyphs becoming
available before it can start.
- Start tracking individual icons needed for a tile, although we don't
do anything with the information yet.
- Introduce typedef for GlyphID