* [core] geometry.hpp 0.8.0
* geojsonvt @ 6.0.0
* Update platform deps, build scripts
* Perf optimizations/cleanup
* Rebase in geometry@080
* D.R.Y. etc
* Ensure fill annotation geometries have closed rings.
* Optimizations
* Update to geojsonvt @ 6.1.0 for clean handoff between geojson parsing and geojsonvt
* Apply close multi/poly geoms for line annotations as well
The current ::stop() implementation was not thread safe and we
should queue the ::stop() for the next loop iteration so even
if we call ::stop() before ::run(), it gets processed.
Not supported, because we don't have control over the Looper
timeout like we do for Loopers we create. On the main thread,
Android takes care of making the application sleep and sets
an arbitrary timeout. This could be properly fixed by implementing
timers in Java, but we don't really have the use case of timers
in the main thread.
[android] #5104 - make nativeLatLngForPix faster (52ms - 45ms).
[andriod] #5104 - make nativeSetLatLng faster (117ms to 105ms)
[android] #5104 - make nativeFlyTo faster (106ms to 97ms)
[android] #5104 - make nativeEaseTo faster (103 to 94ms)
[android] #5104 - make nativeJumpTo faster (94 to 86ms)
[android] #5104 - make nativeProjectedMetersForLatLng faster (32 to 30ms)
[android] #5104 - make nativeLatLngForProjectedMeters faster (32 to 30ms)
* correct all EXPECT_EQ(actual, expected) to EXPECT_EQ(expected, actual)
* fix getMapboxURLPathname() of URL with querystring
* add test for normalizeSourceURL of non-mapbox protocal
* Update mapbox.cpp
- Use glBindAttribLocation for GLSL attributes.
- Create a separate shader for each shader that supports overdraw.
Needed because each uniform location must be known for every program.
- Create a separate VAO for each shader inside buckets.
Needed because we can only bind a VAO to a specific shader.
Fixes#5435.
- puts function definitions in a namespace ... {} rather than using namespace ...;
- remove trailing whitespace
- add trailing newline
- protect SQL statements from being formatted by clang-format
32 bit integers should be enough for zoom scale logic. In shape
annotation logic, 'maxAmountOfTileFeatures' requires 64 bits because we
are multiplying the zoom scale with the extent, which might give a
number higher than std::numeric_limits<uint32_t>::max().
