The newly introduced `GeometryTile::reset()` is
used while GeoJSON tile update in order to prevent from the
parsing of the new data with the stale layers or vice verse.
- share the `GeoJSONOptions` instances using `Immutable<GeoJSONOptions>` - avoid extra copying
- fix wrapping of the `GeoJSONOptions` instances in supercluster map/reduce lambdas.
Previously, local variables were wrapped by reference.
The newly introduced `Scheduler::GetSequenced()` returns
sequenced schedulers from the cache limited to 10 instances,
preventing from spawning too many threads.
In https://github.com/mapbox/mapbox-gl-native/pull/15862, we introduced individual textures for line dash patterns to eliminate atlas overflows. Unfortunately, this broke dashed lines that had round caps (dashed lines with straight caps still rendered correctly). Line pattern textures for round caps were now using 256×15 pixel textures.
The OpenGL ES 2.0 spec, section 3.8.2 states:
Calling a sampler from a fragment shader will return `(R,G,B,A) = (0,0,0,1)` if any of the following conditions are true:
[…]
- A two-dimensional sampler is called, the corresponding texture image is a non-power-of-two image […], and either the texture wrap mode is not `CLAMP_TO_EDGE`, or the minification filter is neither `NEAREST` nor `LINEAR`.
[…]
This means that texture lookups won't work for NPOT textures unless they use `GL_CLAMP_TO_EDGE`. We're using `GL_CLAMP_TO_EDGE` for the vertical direction, but GL_REPEAT for the horizontal direction, which means that we need a power-of-two texture for our line dash patterns to work on OpenGL ES 2.0 conforming implementations. Fortunately, this just means changing the height from 15 to 16, and from 30 to 32, so we don't waste many pixels.
MSVC implementation of std::promise is buggy and only works with types
that can be default-constructed. To avoid a compilation failure in the
instantiation of ask() inside MapSnapshotter::getRegion(), which creates
a std::promise<LanLngBounds>, make LatLngBounds' default constructor
public.
Before this change, all GeoJSON source cleared tile pyramid
at the style transition and it caused flickering.
Now, in Continuous mode, we keep the existing tiles until the
new GeoJSON data is not yet available, thus providing smart
style transitions without flickering.
In other modes (Static, Tile) clear the tile pyramid in order
to avoid render tests being flaky.
* [core] Bump gl-js version
* [core] Implement image expression
* [core] Use new image expression
* [core] Coerce image expression to / from string
* [core] Serialize evaluated image
* [core] Pass available images to layout
* [core] Pass images to evaluation context
* [core] Set available flag value based on image availability
* [core] Allow image coercion to boolean to indicate image availability
* [core] Coalesce image expression
* [core] Add image expression to next build system
* [core] Align serialization format and evaluated type with gl-js
* [core] Add images to expression evaluation method
* [core] Add support for Image expression to expression test runner
* [core] Unskip image expression tests
* [core] Update unit tests
* [core] Use image expression in annotation manager
* [core] Add string to ImageExpression conversion
* [core] Add image expression to expression dsl
* [core] Convert tokens for implicitly created Image literal
* [core] Fix clang format
* [core] Split generated style code lines that are over 120 characters
* [core] Add unit test for image expression equality
* [core] Add image property expression evaluation unit test
* [core] Unskip image expression render test
* [core] Skip 'in' expression tests
* [core] Ignore fill-pattern/update-feature-state render test
* [core] Rename Image::serialize to Image::toValue
Composing tiles from the GeoJSON features is an expensive
operation that might block UI thread on updating the `GeoJsonSource`
with the new data.
This change moves tile composing to the background thread and thus
unblocks the UI thread.
This moves the LineAtlas from a shared texture that contained SDF dash patterns to use individual textures.
Previously, the texture space was limited to a texture of 512 pixels height. Dash patterns were never removed (and are still never removed as of this patch), which means that this texture could fill up for styles that use a lot of different dash patterns. In particular, dash patterns for lines with a round line cap take up 15 pixels of texture height, limiting the amount of unique dash patterns to 34. While this was probably enough for rendering a single style, we quickly exhausted this number when reusing the Map object to render different styles.
Instead of a global shared texture, we're now creating individual textures for every dash pattern. These textures are still cached so that we don't need to re-upload the texture on every frame.
This fixes rendering by account for the 1px texture padding around icons that were stretched with icon-text-fit. We've added the 1px padding before, but didn't scale it accordingly when we are resizing the icon when it is stretched to fit the text. Adjusts the code to match the logic in GL JS.
Unlike icons, pattern changes are not caught with
style-diff meaning that the existing request could
be from the previous style and we cannot coalesce
requests for them.
This commit refactors `utils::ThreadPool` into a template
`ThreadedScheduler` class and provides aux type aliases.
So that it is possible to obtain a sequenced schedule,where
all the scheduled tasks are guarantied to be executed
consequently.
The sequenced lightweight scheduler is required by both the
orchestration thread and the refactored `FileSource` implementation.
RenderOrchestrator holds a vector of ordered rendered layers,
in order to minimize calling of `getRenderLayer()`, which
could be quite expensive as it calculates hash of the layer Id.
This change improves orchestration performance in particular
for the styles that have multiple sources.
The `API_renderStill_multiple_sources` benchmark runs
approximately 70% faster with the changes applied.
