This argument is now non optional, but this never hits the same bad access.
I voted to simplify the code here since the argument is never used optionally in our codebase.
Some Macs encounter an internal error when compiling cluster_render.glsl
caused by a likely bug in the MVK/Metal
compiler when using gl_HelperInvocation.
1. Viewport::get_visible_rect().position is always zero.
So Control::get_window_rect is identical to Control::get_global_rect.
Remove Control::get_window_rect since it is not used in the source code.
2. sqrt(a * a) = abs(a) for doubles
3. Simplify affine_inverse combination
4. Simplify calculation in shaders
Previously, only forward basis distance from the camera was used.
This means that unnecessarily high LOD levels were used for objects located to the side of the camera.
The distance from the camera origin is now used, independently of direction.
Replace all TODO uses of `#warning` by proper TODO comments, and will open
matching bug reports to keep track of them.
We don't have a great track record fixing TODOs, but I'd wager we're even
worse for fixing these "TODO #warning" so we should prohibit this usage.
Negatively scaled objects should be mirrored. This is already implemented, but it breaks when mirrored and non-mirrored instances of the same object are visible together.
It turns out that the code that skips-over repeats in `RenderForwardClustered::_render_list_template` also skips the code that accounts for the culling mode of mirrored objects.
The solution here is to consider the `mirror` flag when determining repeats. This might result in more draw commands than necessary since a mirrored object can split a group of non-mirrored instances in two.
This problem doesn't appear in the mobile renderer because the repeat optimization isn't implemented there yet.
The problem still appears in MultiMeshInstance3D in *all* renderers.
Fixes#62879 and #58546.
Mainly:
- Make `max_descriptors_per_pool` project setting Vulkan-specific.
- Use a common, render driver agnostic magic FourCC for shader binary data.
- Downgrade spirv_reflect to Vulkan-only dependency.
- Add a `RENDER_DRIVER_*` macro to GLSL shader code for per-driver customizations.
This removes the countless small UBO writes we had before
and replaces them with a single large write per render pass.
This results in much faster rendering on low-end devices
but improves speed on all devices.
