Was tripping up Xcode 11 betas with:
Imported declaration 'UITableViewDiffableDataSourceCellProvider' could not be mapped to 'UITableViewDiffableDataSourceReference.CellProvider'
Do not flood the RunLoop with wake up calls, potentially causing the
socket buffer to overflow.
Also, write calls are syscalls and will cause a context switch.
Follow the approach from mapbox-gl-js for enabling opaque pass for background layer (same as for fill layer).
Fix combinations/heatmap-translucent--background-opaque render test.
We were capturing the reference to an object that could go away when the request was canceled. However, the lambda could still be invoked by NSURLSession, creating a race condition with a resulting crash when the response to a tile request was 404 Not Found.
Patch partly fixes#15163 in a way that it doesn't allow loading tens of thousands of tiles and attempt to show area above horizon:
Limit pitch based on edge insets. It is not too bad - current limit of 60 degrees stays active until center of perspective is moved towards the bottom, to 84% of screen height. The plan is to split removal of 60 degrees limit to follow up patch.
Fix max Z calculation in getProjMatrix. TransformState::getProjMatrix calculation of farZ was complex with possibility to lead to negative z values. Replacing it with simpler, precise calculation:
furthestDistance = cameraToCenterDistance / (1 - tanFovAboveCenter * std::tan(getPitch()));
TransformState::getProjMatrix calculation of farZ was an aproximation. Replacing it with simpler, but precise calculation.
Related to: #15163
Camera/Render mode updates can be streamed from various different stateless sources and we should make sure to skip an update to a mode that's already set. This keeps the callbacks tidy and does not perform unnecessary work.
When all location layer animations were reset, due to the mode change, we were recalculating the time of the animation, while we should be taking the remaining duration. Frequent recalculation requests resulted in a really short animation duration.
Viewport center offset usage was wrongly submitted in #14664. It was part of alternative approach that used enlarged viewport. Existing and added tests were not sufficient to spot the regression, since the collision check padding is usually larger than the center offset x and y. Annotation picking has tolerance of only 10 pixels but no annotation integration test was using content insets.
Usage of offset is not needed because `posMatrix` in e.g. `CollisionIndex::projectPoint(const mat4& posMatrix, const Point<float>& point)` already incorporates center offset (projection matrix) and the code in current master was just offsetting all by the value.
Modified [ios] MGLAnnotationViewIntegrationTests.testSelectingAnnotationWithCenterOffset to use different insets. It verifies the fix.
Fixes [iOS] Annotations are not selectable (added via iosapp menu) #15106:
In case of #15106, view's original content insets is {top:88, bottom:34}, causing that center offset is {x:0, y:27} and selection with tolerance of 10 wouldn't select annotation.
After tapping the view, so that the header gets removed, view's content insets get changed to {top:44, bottom:34}, center offset is {x:0, y:5} and annotation selection would work, as described in #15106.
Fixes: #15106
