The implementation is buggy and not worth maintaining anymore
because performance benefits are not substantial or sometimes
worse. Also, removing it saves about 150 ~ 180 KB in binary size.
Below timings are averages of minimum 5 runs.
```
Device Init launch Average relaunch
s10 1129.8 ms 700 ms
s10 - binary 1346.75 ms 694 ms
Pixel 1692 ms 723 ms
Pixel - binary 1883 ms 1039 ms
Kazam 17948 ms 1339 ms
Kazam - binary 19157 ms 1564 ms
Wiko 2060 ms 1278 ms
Wiko - binary 3876 ms 1136 ms
```
Fixes#14294
Remove re-linking programs as redundant.
It costs (cheaper to link once than twice) and (subjective) is not that common GL API
usage pattern, although perfectly legal and permitted.
Initial idea, behind the removed code, was to enable work on optimization
that would reduce number of attrib setup calls in case when VAO is not
available (as described in #9433).
As such optimization is not implemented, and it is arguable if it makes sense
to do it now, we can remove re-linking.
Related to closed PRs #9433 and PR #11583.
I have [measured the time spent just on relinking](https://gist.github.com/astojilj/29bd5a5c5dc0b2d9f29ecb660da07fbf) using release build on iPhone SE (A9, same as iPhone 6S):
- 1st run after reboot or installation
Total 37.14ms, average per program:1.86ms
- reopening
Total: 2.47ms, average per program: 0.12ms
This time we save using the patch here.
A RenderTile is already never created for a not renderable tile, guarantied by the checks in `updateRenderables()`.
However, the client code had plenty of `isRenderable()` checks in the render path, which complicated the code and affected rendering performance.
This patch removes the unneeded checks from the client code and puts an assertion to `TilePyramid::addRenderTile()`.
coordinatePointMatrix is always used with getZoom():
Projection::worldSize(scale) / std::pow(2, z) is computed as
scale * util::tileSize / std::pow(2, z)
scale is equal to std::pow(2, z).
____________________________________=>
Projection::worldSize(scale) / std::pow(2, z) => util::tileSize.
- Do not carry it over everywhere as parameter, it is a shared
instance anyway and the lifecycle is pretty much the app lifecycle
from the moment we instantiate a map.
- Rename to BackgroundScheduler because it is a Scheduler that will
do tasks in the background, we don't make assumptions if it is a
thread pool or a single thread.
- Most importantly, remove the dependency from `core` on `platform`.
Before this change, ImageManger's 'done' callback for onStyleImageMissing observer
notification that was created on renderer thread, could be called from different thread,
therefore, is not thread safe. For example, on Android platform, callback was invoked
from UI thread. This change makes callback to be scheduled on originating thread.
Symbols are placed accordingly to their viewport Y order,
if the style `symbol-z-order` is set to `viewport-y`.
This improves rendering of symbol layers, where icons are allowed
to overlap but not text.
JSON keys in our style specification don't have a defined order. This change sorts them alphabetically so that we can rely on the order remaining them same across code generation runs.
