use bevy_app::Plugin;
use bevy_asset::{load_internal_asset, AssetId, Handle};

use bevy_core_pipeline::{
    core_2d::{AlphaMask2d, Camera2d, Opaque2d, Transparent2d, CORE_2D_DEPTH_FORMAT},
    tonemapping::{
        get_lut_bind_group_layout_entries, get_lut_bindings, Tonemapping, TonemappingLuts,
    },
};
use bevy_derive::{Deref, DerefMut};
use bevy_ecs::{
    entity::EntityHashMap,
    prelude::*,
    query::ROQueryItem,
    system::{lifetimeless::*, SystemParamItem, SystemState},
};
use bevy_math::{Affine3, Vec4};
use bevy_reflect::{std_traits::ReflectDefault, Reflect};
use bevy_render::batching::gpu_preprocessing::IndirectParameters;
use bevy_render::batching::no_gpu_preprocessing::batch_and_prepare_binned_render_phase;
use bevy_render::batching::no_gpu_preprocessing::{
    self, batch_and_prepare_sorted_render_phase, write_batched_instance_buffer,
    BatchedInstanceBuffer,
};
use bevy_render::batching::GetFullBatchData;
use bevy_render::mesh::allocator::MeshAllocator;
use bevy_render::mesh::{MeshVertexBufferLayoutRef, RenderMesh};
use bevy_render::texture::FallbackImage;
use bevy_render::{
    batching::{GetBatchData, NoAutomaticBatching},
    globals::{GlobalsBuffer, GlobalsUniform},
    mesh::{Mesh, RenderMeshBufferInfo},
    render_asset::RenderAssets,
    render_phase::{PhaseItem, RenderCommand, RenderCommandResult, TrackedRenderPass},
    render_resource::{binding_types::uniform_buffer, *},
    renderer::{RenderDevice, RenderQueue},
    texture::{
        BevyDefault, DefaultImageSampler, GpuImage, Image, ImageSampler, TextureFormatPixelInfo,
    },
    view::{
        ExtractedView, ViewTarget, ViewUniform, ViewUniformOffset, ViewUniforms, ViewVisibility,
    },
    Extract, ExtractSchedule, Render, RenderApp, RenderSet,
};
use bevy_transform::components::GlobalTransform;
use bevy_utils::tracing::error;
use nonmax::NonMaxU32;

use crate::Material2dBindGroupId;

/// Component for rendering with meshes in the 2d pipeline, usually with a [2d material](crate::Material2d) such as [`ColorMaterial`](crate::ColorMaterial).
///
/// It wraps a [`Handle<Mesh>`] to differentiate from the 3d pipelines which use the handles directly as components
#[derive(Default, Clone, Component, Debug, Reflect, PartialEq, Eq, Deref, DerefMut)]
#[reflect(Default, Component)]
pub struct Mesh2dHandle(pub Handle<Mesh>);

impl From<Handle<Mesh>> for Mesh2dHandle {
    fn from(handle: Handle<Mesh>) -> Self {
        Self(handle)
    }
}

#[derive(Default)]
pub struct Mesh2dRenderPlugin;

pub const MESH2D_VERTEX_OUTPUT: Handle<Shader> = Handle::weak_from_u128(7646632476603252194);
pub const MESH2D_VIEW_TYPES_HANDLE: Handle<Shader> = Handle::weak_from_u128(12677582416765805110);
pub const MESH2D_VIEW_BINDINGS_HANDLE: Handle<Shader> = Handle::weak_from_u128(6901431444735842434);
pub const MESH2D_TYPES_HANDLE: Handle<Shader> = Handle::weak_from_u128(8994673400261890424);
pub const MESH2D_BINDINGS_HANDLE: Handle<Shader> = Handle::weak_from_u128(8983617858458862856);
pub const MESH2D_FUNCTIONS_HANDLE: Handle<Shader> = Handle::weak_from_u128(4976379308250389413);
pub const MESH2D_SHADER_HANDLE: Handle<Shader> = Handle::weak_from_u128(2971387252468633715);

impl Plugin for Mesh2dRenderPlugin {
    fn build(&self, app: &mut bevy_app::App) {
        load_internal_asset!(
            app,
            MESH2D_VERTEX_OUTPUT,
            "mesh2d_vertex_output.wgsl",
            Shader::from_wgsl
        );
        load_internal_asset!(
            app,
            MESH2D_VIEW_TYPES_HANDLE,
            "mesh2d_view_types.wgsl",
            Shader::from_wgsl
        );
        load_internal_asset!(
            app,
            MESH2D_VIEW_BINDINGS_HANDLE,
            "mesh2d_view_bindings.wgsl",
            Shader::from_wgsl
        );
        load_internal_asset!(
            app,
            MESH2D_TYPES_HANDLE,
            "mesh2d_types.wgsl",
            Shader::from_wgsl
        );
        load_internal_asset!(
            app,
            MESH2D_FUNCTIONS_HANDLE,
            "mesh2d_functions.wgsl",
            Shader::from_wgsl
        );
        load_internal_asset!(app, MESH2D_SHADER_HANDLE, "mesh2d.wgsl", Shader::from_wgsl);

        if let Some(render_app) = app.get_sub_app_mut(RenderApp) {
            render_app
                .init_resource::<RenderMesh2dInstances>()
                .init_resource::<SpecializedMeshPipelines<Mesh2dPipeline>>()
                .add_systems(ExtractSchedule, extract_mesh2d)
                .add_systems(
                    Render,
                    (
                        batch_and_prepare_binned_render_phase::<Opaque2d, Mesh2dPipeline>
                            .in_set(RenderSet::PrepareResources),
                        batch_and_prepare_binned_render_phase::<AlphaMask2d, Mesh2dPipeline>
                            .in_set(RenderSet::PrepareResources),
                        batch_and_prepare_sorted_render_phase::<Transparent2d, Mesh2dPipeline>
                            .in_set(RenderSet::PrepareResources),
                        write_batched_instance_buffer::<Mesh2dPipeline>
                            .in_set(RenderSet::PrepareResourcesFlush),
                        prepare_mesh2d_bind_group.in_set(RenderSet::PrepareBindGroups),
                        prepare_mesh2d_view_bind_groups.in_set(RenderSet::PrepareBindGroups),
                        no_gpu_preprocessing::clear_batched_cpu_instance_buffers::<Mesh2dPipeline>
                            .in_set(RenderSet::Cleanup)
                            .after(RenderSet::Render),
                    ),
                );
        }
    }

    fn finish(&self, app: &mut bevy_app::App) {
        let mut mesh_bindings_shader_defs = Vec::with_capacity(1);

        if let Some(render_app) = app.get_sub_app_mut(RenderApp) {
            let render_device = render_app.world().resource::<RenderDevice>();
            let batched_instance_buffer =
                BatchedInstanceBuffer::<Mesh2dUniform>::new(render_device);

            if let Some(per_object_buffer_batch_size) =
                GpuArrayBuffer::<Mesh2dUniform>::batch_size(render_device)
            {
                mesh_bindings_shader_defs.push(ShaderDefVal::UInt(
                    "PER_OBJECT_BUFFER_BATCH_SIZE".into(),
                    per_object_buffer_batch_size,
                ));
            }

            render_app
                .insert_resource(batched_instance_buffer)
                .init_resource::<Mesh2dPipeline>();
        }

        // Load the mesh_bindings shader module here as it depends on runtime information about
        // whether storage buffers are supported, or the maximum uniform buffer binding size.
        load_internal_asset!(
            app,
            MESH2D_BINDINGS_HANDLE,
            "mesh2d_bindings.wgsl",
            Shader::from_wgsl_with_defs,
            mesh_bindings_shader_defs
        );
    }
}

#[derive(Component)]
pub struct Mesh2dTransforms {
    pub world_from_local: Affine3,
    pub flags: u32,
}

#[derive(ShaderType, Clone, Copy)]
pub struct Mesh2dUniform {
    // Affine 4x3 matrix transposed to 3x4
    pub world_from_local: [Vec4; 3],
    // 3x3 matrix packed in mat2x4 and f32 as:
    //   [0].xyz, [1].x,
    //   [1].yz, [2].xy
    //   [2].z
    pub local_from_world_transpose_a: [Vec4; 2],
    pub local_from_world_transpose_b: f32,
    pub flags: u32,
}

impl From<&Mesh2dTransforms> for Mesh2dUniform {
    fn from(mesh_transforms: &Mesh2dTransforms) -> Self {
        let (local_from_world_transpose_a, local_from_world_transpose_b) =
            mesh_transforms.world_from_local.inverse_transpose_3x3();
        Self {
            world_from_local: mesh_transforms.world_from_local.to_transpose(),
            local_from_world_transpose_a,
            local_from_world_transpose_b,
            flags: mesh_transforms.flags,
        }
    }
}

// NOTE: These must match the bit flags in bevy_sprite/src/mesh2d/mesh2d.wgsl!
bitflags::bitflags! {
    #[repr(transparent)]
    pub struct MeshFlags: u32 {
        const NONE                       = 0;
        const UNINITIALIZED              = 0xFFFF;
    }
}

pub struct RenderMesh2dInstance {
    pub transforms: Mesh2dTransforms,
    pub mesh_asset_id: AssetId<Mesh>,
    pub material_bind_group_id: Material2dBindGroupId,
    pub automatic_batching: bool,
}

#[derive(Default, Resource, Deref, DerefMut)]
pub struct RenderMesh2dInstances(EntityHashMap<RenderMesh2dInstance>);

#[derive(Component)]
pub struct Mesh2d;

pub fn extract_mesh2d(
    mut commands: Commands,
    mut previous_len: Local<usize>,
    mut render_mesh_instances: ResMut<RenderMesh2dInstances>,
    query: Extract<
        Query<(
            Entity,
            &ViewVisibility,
            &GlobalTransform,
            &Mesh2dHandle,
            Has<NoAutomaticBatching>,
        )>,
    >,
) {
    render_mesh_instances.clear();
    let mut entities = Vec::with_capacity(*previous_len);

    for (entity, view_visibility, transform, handle, no_automatic_batching) in &query {
        if !view_visibility.get() {
            continue;
        }
        // FIXME: Remove this - it is just a workaround to enable rendering to work as
        // render commands require an entity to exist at the moment.
        entities.push((entity, Mesh2d));
        render_mesh_instances.insert(
            entity,
            RenderMesh2dInstance {
                transforms: Mesh2dTransforms {
                    world_from_local: (&transform.affine()).into(),
                    flags: MeshFlags::empty().bits(),
                },
                mesh_asset_id: handle.0.id(),
                material_bind_group_id: Material2dBindGroupId::default(),
                automatic_batching: !no_automatic_batching,
            },
        );
    }
    *previous_len = entities.len();
    commands.insert_or_spawn_batch(entities);
}

#[derive(Resource, Clone)]
pub struct Mesh2dPipeline {
    pub view_layout: BindGroupLayout,
    pub mesh_layout: BindGroupLayout,
    // This dummy white texture is to be used in place of optional textures
    pub dummy_white_gpu_image: GpuImage,
    pub per_object_buffer_batch_size: Option<u32>,
}

impl FromWorld for Mesh2dPipeline {
    fn from_world(world: &mut World) -> Self {
        let mut system_state: SystemState<(
            Res<RenderDevice>,
            Res<RenderQueue>,
            Res<DefaultImageSampler>,
        )> = SystemState::new(world);
        let (render_device, render_queue, default_sampler) = system_state.get_mut(world);
        let render_device = render_device.into_inner();
        let tonemapping_lut_entries = get_lut_bind_group_layout_entries();
        let view_layout = render_device.create_bind_group_layout(
            "mesh2d_view_layout",
            &BindGroupLayoutEntries::with_indices(
                ShaderStages::VERTEX_FRAGMENT,
                (
                    (0, uniform_buffer::<ViewUniform>(true)),
                    (1, uniform_buffer::<GlobalsUniform>(false)),
                    (
                        2,
                        tonemapping_lut_entries[0].visibility(ShaderStages::FRAGMENT),
                    ),
                    (
                        3,
                        tonemapping_lut_entries[1].visibility(ShaderStages::FRAGMENT),
                    ),
                ),
            ),
        );

        let mesh_layout = render_device.create_bind_group_layout(
            "mesh2d_layout",
            &BindGroupLayoutEntries::single(
                ShaderStages::VERTEX_FRAGMENT,
                GpuArrayBuffer::<Mesh2dUniform>::binding_layout(render_device),
            ),
        );
        // A 1x1x1 'all 1.0' texture to use as a dummy texture to use in place of optional StandardMaterial textures
        let dummy_white_gpu_image = {
            let image = Image::default();
            let texture = render_device.create_texture(&image.texture_descriptor);
            let sampler = match image.sampler {
                ImageSampler::Default => (**default_sampler).clone(),
                ImageSampler::Descriptor(ref descriptor) => {
                    render_device.create_sampler(&descriptor.as_wgpu())
                }
            };

            let format_size = image.texture_descriptor.format.pixel_size();
            render_queue.write_texture(
                texture.as_image_copy(),
                &image.data,
                ImageDataLayout {
                    offset: 0,
                    bytes_per_row: Some(image.width() * format_size as u32),
                    rows_per_image: None,
                },
                image.texture_descriptor.size,
            );

            let texture_view = texture.create_view(&TextureViewDescriptor::default());
            GpuImage {
                texture,
                texture_view,
                texture_format: image.texture_descriptor.format,
                sampler,
                size: image.size(),
                mip_level_count: image.texture_descriptor.mip_level_count,
            }
        };
        Mesh2dPipeline {
            view_layout,
            mesh_layout,
            dummy_white_gpu_image,
            per_object_buffer_batch_size: GpuArrayBuffer::<Mesh2dUniform>::batch_size(
                render_device,
            ),
        }
    }
}

impl Mesh2dPipeline {
    pub fn get_image_texture<'a>(
        &'a self,
        gpu_images: &'a RenderAssets<GpuImage>,
        handle_option: &Option<Handle<Image>>,
    ) -> Option<(&'a TextureView, &'a Sampler)> {
        if let Some(handle) = handle_option {
            let gpu_image = gpu_images.get(handle)?;
            Some((&gpu_image.texture_view, &gpu_image.sampler))
        } else {
            Some((
                &self.dummy_white_gpu_image.texture_view,
                &self.dummy_white_gpu_image.sampler,
            ))
        }
    }
}

impl GetBatchData for Mesh2dPipeline {
    type Param = (
        SRes<RenderMesh2dInstances>,
        SRes<RenderAssets<RenderMesh>>,
        SRes<MeshAllocator>,
    );
    type CompareData = (Material2dBindGroupId, AssetId<Mesh>);
    type BufferData = Mesh2dUniform;

    fn get_batch_data(
        (mesh_instances, _, _): &SystemParamItem<Self::Param>,
        entity: Entity,
    ) -> Option<(Self::BufferData, Option<Self::CompareData>)> {
        let mesh_instance = mesh_instances.get(&entity)?;
        Some((
            (&mesh_instance.transforms).into(),
            mesh_instance.automatic_batching.then_some((
                mesh_instance.material_bind_group_id,
                mesh_instance.mesh_asset_id,
            )),
        ))
    }
}

impl GetFullBatchData for Mesh2dPipeline {
    type BufferInputData = ();

    fn get_binned_batch_data(
        (mesh_instances, _, _): &SystemParamItem<Self::Param>,
        entity: Entity,
    ) -> Option<Self::BufferData> {
        let mesh_instance = mesh_instances.get(&entity)?;
        Some((&mesh_instance.transforms).into())
    }

    fn get_index_and_compare_data(
        _: &SystemParamItem<Self::Param>,
        _query_item: Entity,
    ) -> Option<(NonMaxU32, Option<Self::CompareData>)> {
        error!(
            "`get_index_and_compare_data` is only intended for GPU mesh uniform building, \
            but this is not yet implemented for 2d meshes"
        );
        None
    }

    fn get_binned_index(
        _: &SystemParamItem<Self::Param>,
        _query_item: Entity,
    ) -> Option<NonMaxU32> {
        error!(
            "`get_binned_index` is only intended for GPU mesh uniform building, \
            but this is not yet implemented for 2d meshes"
        );
        None
    }

    fn get_batch_indirect_parameters_index(
        (mesh_instances, meshes, mesh_allocator): &SystemParamItem<Self::Param>,
        indirect_parameters_buffer: &mut bevy_render::batching::gpu_preprocessing::IndirectParametersBuffer,
        entity: Entity,
        instance_index: u32,
    ) -> Option<NonMaxU32> {
        let mesh_instance = mesh_instances.get(&entity)?;
        let mesh = meshes.get(mesh_instance.mesh_asset_id)?;
        let vertex_buffer_slice = mesh_allocator.mesh_vertex_slice(&mesh_instance.mesh_asset_id)?;

        // Note that `IndirectParameters` covers both of these structures, even
        // though they actually have distinct layouts. See the comment above that
        // type for more information.
        let indirect_parameters = match mesh.buffer_info {
            RenderMeshBufferInfo::Indexed {
                count: index_count, ..
            } => {
                let index_buffer_slice =
                    mesh_allocator.mesh_index_slice(&mesh_instance.mesh_asset_id)?;
                IndirectParameters {
                    vertex_or_index_count: index_count,
                    instance_count: 0,
                    first_vertex_or_first_index: index_buffer_slice.range.start,
                    base_vertex_or_first_instance: vertex_buffer_slice.range.start,
                    first_instance: instance_index,
                }
            }
            RenderMeshBufferInfo::NonIndexed => IndirectParameters {
                vertex_or_index_count: mesh.vertex_count,
                instance_count: 0,
                first_vertex_or_first_index: vertex_buffer_slice.range.start,
                base_vertex_or_first_instance: instance_index,
                first_instance: instance_index,
            },
        };

        (indirect_parameters_buffer.push(indirect_parameters) as u32)
            .try_into()
            .ok()
    }
}

bitflags::bitflags! {
    #[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)]
    #[repr(transparent)]
    // NOTE: Apparently quadro drivers support up to 64x MSAA.
    // MSAA uses the highest 3 bits for the MSAA log2(sample count) to support up to 128x MSAA.
    // FIXME: make normals optional?
    pub struct Mesh2dPipelineKey: u32 {
        const NONE                              = 0;
        const HDR                               = 1 << 0;
        const TONEMAP_IN_SHADER                 = 1 << 1;
        const DEBAND_DITHER                     = 1 << 2;
        const BLEND_ALPHA                       = 1 << 3;
        const MAY_DISCARD                       = 1 << 4;
        const MSAA_RESERVED_BITS                = Self::MSAA_MASK_BITS << Self::MSAA_SHIFT_BITS;
        const PRIMITIVE_TOPOLOGY_RESERVED_BITS  = Self::PRIMITIVE_TOPOLOGY_MASK_BITS << Self::PRIMITIVE_TOPOLOGY_SHIFT_BITS;
        const TONEMAP_METHOD_RESERVED_BITS      = Self::TONEMAP_METHOD_MASK_BITS << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_NONE               = 0 << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_REINHARD           = 1 << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_REINHARD_LUMINANCE = 2 << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_ACES_FITTED        = 3 << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_AGX                = 4 << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_SOMEWHAT_BORING_DISPLAY_TRANSFORM = 5 << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_TONY_MC_MAPFACE    = 6 << Self::TONEMAP_METHOD_SHIFT_BITS;
        const TONEMAP_METHOD_BLENDER_FILMIC     = 7 << Self::TONEMAP_METHOD_SHIFT_BITS;
    }
}

impl Mesh2dPipelineKey {
    const MSAA_MASK_BITS: u32 = 0b111;
    const MSAA_SHIFT_BITS: u32 = 32 - Self::MSAA_MASK_BITS.count_ones();
    const PRIMITIVE_TOPOLOGY_MASK_BITS: u32 = 0b111;
    const PRIMITIVE_TOPOLOGY_SHIFT_BITS: u32 = Self::MSAA_SHIFT_BITS - 3;
    const TONEMAP_METHOD_MASK_BITS: u32 = 0b111;
    const TONEMAP_METHOD_SHIFT_BITS: u32 =
        Self::PRIMITIVE_TOPOLOGY_SHIFT_BITS - Self::TONEMAP_METHOD_MASK_BITS.count_ones();

    pub fn from_msaa_samples(msaa_samples: u32) -> Self {
        let msaa_bits =
            (msaa_samples.trailing_zeros() & Self::MSAA_MASK_BITS) << Self::MSAA_SHIFT_BITS;
        Self::from_bits_retain(msaa_bits)
    }

    pub fn from_hdr(hdr: bool) -> Self {
        if hdr {
            Mesh2dPipelineKey::HDR
        } else {
            Mesh2dPipelineKey::NONE
        }
    }

    pub fn msaa_samples(&self) -> u32 {
        1 << ((self.bits() >> Self::MSAA_SHIFT_BITS) & Self::MSAA_MASK_BITS)
    }

    pub fn from_primitive_topology(primitive_topology: PrimitiveTopology) -> Self {
        let primitive_topology_bits = ((primitive_topology as u32)
            & Self::PRIMITIVE_TOPOLOGY_MASK_BITS)
            << Self::PRIMITIVE_TOPOLOGY_SHIFT_BITS;
        Self::from_bits_retain(primitive_topology_bits)
    }

    pub fn primitive_topology(&self) -> PrimitiveTopology {
        let primitive_topology_bits = (self.bits() >> Self::PRIMITIVE_TOPOLOGY_SHIFT_BITS)
            & Self::PRIMITIVE_TOPOLOGY_MASK_BITS;
        match primitive_topology_bits {
            x if x == PrimitiveTopology::PointList as u32 => PrimitiveTopology::PointList,
            x if x == PrimitiveTopology::LineList as u32 => PrimitiveTopology::LineList,
            x if x == PrimitiveTopology::LineStrip as u32 => PrimitiveTopology::LineStrip,
            x if x == PrimitiveTopology::TriangleList as u32 => PrimitiveTopology::TriangleList,
            x if x == PrimitiveTopology::TriangleStrip as u32 => PrimitiveTopology::TriangleStrip,
            _ => PrimitiveTopology::default(),
        }
    }
}

impl SpecializedMeshPipeline for Mesh2dPipeline {
    type Key = Mesh2dPipelineKey;

    fn specialize(
        &self,
        key: Self::Key,
        layout: &MeshVertexBufferLayoutRef,
    ) -> Result<RenderPipelineDescriptor, SpecializedMeshPipelineError> {
        let mut shader_defs = Vec::new();
        let mut vertex_attributes = Vec::new();

        if layout.0.contains(Mesh::ATTRIBUTE_POSITION) {
            shader_defs.push("VERTEX_POSITIONS".into());
            vertex_attributes.push(Mesh::ATTRIBUTE_POSITION.at_shader_location(0));
        }

        if layout.0.contains(Mesh::ATTRIBUTE_NORMAL) {
            shader_defs.push("VERTEX_NORMALS".into());
            vertex_attributes.push(Mesh::ATTRIBUTE_NORMAL.at_shader_location(1));
        }

        if layout.0.contains(Mesh::ATTRIBUTE_UV_0) {
            shader_defs.push("VERTEX_UVS".into());
            vertex_attributes.push(Mesh::ATTRIBUTE_UV_0.at_shader_location(2));
        }

        if layout.0.contains(Mesh::ATTRIBUTE_TANGENT) {
            shader_defs.push("VERTEX_TANGENTS".into());
            vertex_attributes.push(Mesh::ATTRIBUTE_TANGENT.at_shader_location(3));
        }

        if layout.0.contains(Mesh::ATTRIBUTE_COLOR) {
            shader_defs.push("VERTEX_COLORS".into());
            vertex_attributes.push(Mesh::ATTRIBUTE_COLOR.at_shader_location(4));
        }

        if key.contains(Mesh2dPipelineKey::TONEMAP_IN_SHADER) {
            shader_defs.push("TONEMAP_IN_SHADER".into());
            shader_defs.push(ShaderDefVal::UInt(
                "TONEMAPPING_LUT_TEXTURE_BINDING_INDEX".into(),
                2,
            ));
            shader_defs.push(ShaderDefVal::UInt(
                "TONEMAPPING_LUT_SAMPLER_BINDING_INDEX".into(),
                3,
            ));

            let method = key.intersection(Mesh2dPipelineKey::TONEMAP_METHOD_RESERVED_BITS);

            match method {
                Mesh2dPipelineKey::TONEMAP_METHOD_NONE => {
                    shader_defs.push("TONEMAP_METHOD_NONE".into());
                }
                Mesh2dPipelineKey::TONEMAP_METHOD_REINHARD => {
                    shader_defs.push("TONEMAP_METHOD_REINHARD".into());
                }
                Mesh2dPipelineKey::TONEMAP_METHOD_REINHARD_LUMINANCE => {
                    shader_defs.push("TONEMAP_METHOD_REINHARD_LUMINANCE".into());
                }
                Mesh2dPipelineKey::TONEMAP_METHOD_ACES_FITTED => {
                    shader_defs.push("TONEMAP_METHOD_ACES_FITTED".into());
                }
                Mesh2dPipelineKey::TONEMAP_METHOD_AGX => {
                    shader_defs.push("TONEMAP_METHOD_AGX".into());
                }
                Mesh2dPipelineKey::TONEMAP_METHOD_SOMEWHAT_BORING_DISPLAY_TRANSFORM => {
                    shader_defs.push("TONEMAP_METHOD_SOMEWHAT_BORING_DISPLAY_TRANSFORM".into());
                }
                Mesh2dPipelineKey::TONEMAP_METHOD_BLENDER_FILMIC => {
                    shader_defs.push("TONEMAP_METHOD_BLENDER_FILMIC".into());
                }
                Mesh2dPipelineKey::TONEMAP_METHOD_TONY_MC_MAPFACE => {
                    shader_defs.push("TONEMAP_METHOD_TONY_MC_MAPFACE".into());
                }
                _ => {}
            }
            // Debanding is tied to tonemapping in the shader, cannot run without it.
            if key.contains(Mesh2dPipelineKey::DEBAND_DITHER) {
                shader_defs.push("DEBAND_DITHER".into());
            }
        }

        if key.contains(Mesh2dPipelineKey::MAY_DISCARD) {
            shader_defs.push("MAY_DISCARD".into());
        }

        let vertex_buffer_layout = layout.0.get_layout(&vertex_attributes)?;

        let format = match key.contains(Mesh2dPipelineKey::HDR) {
            true => ViewTarget::TEXTURE_FORMAT_HDR,
            false => TextureFormat::bevy_default(),
        };

        let (depth_write_enabled, label, blend);
        if key.contains(Mesh2dPipelineKey::BLEND_ALPHA) {
            label = "transparent_mesh2d_pipeline";
            blend = Some(BlendState::ALPHA_BLENDING);
            depth_write_enabled = false;
        } else {
            label = "opaque_mesh2d_pipeline";
            blend = None;
            depth_write_enabled = true;
        }

        Ok(RenderPipelineDescriptor {
            vertex: VertexState {
                shader: MESH2D_SHADER_HANDLE,
                entry_point: "vertex".into(),
                shader_defs: shader_defs.clone(),
                buffers: vec![vertex_buffer_layout],
            },
            fragment: Some(FragmentState {
                shader: MESH2D_SHADER_HANDLE,
                shader_defs,
                entry_point: "fragment".into(),
                targets: vec![Some(ColorTargetState {
                    format,
                    blend,
                    write_mask: ColorWrites::ALL,
                })],
            }),
            layout: vec![self.view_layout.clone(), self.mesh_layout.clone()],
            push_constant_ranges: vec![],
            primitive: PrimitiveState {
                front_face: FrontFace::Ccw,
                cull_mode: None,
                unclipped_depth: false,
                polygon_mode: PolygonMode::Fill,
                conservative: false,
                topology: key.primitive_topology(),
                strip_index_format: None,
            },
            depth_stencil: Some(DepthStencilState {
                format: CORE_2D_DEPTH_FORMAT,
                depth_write_enabled,
                depth_compare: CompareFunction::GreaterEqual,
                stencil: StencilState {
                    front: StencilFaceState::IGNORE,
                    back: StencilFaceState::IGNORE,
                    read_mask: 0,
                    write_mask: 0,
                },
                bias: DepthBiasState {
                    constant: 0,
                    slope_scale: 0.0,
                    clamp: 0.0,
                },
            }),
            multisample: MultisampleState {
                count: key.msaa_samples(),
                mask: !0,
                alpha_to_coverage_enabled: false,
            },
            label: Some(label.into()),
        })
    }
}

#[derive(Resource)]
pub struct Mesh2dBindGroup {
    pub value: BindGroup,
}

pub fn prepare_mesh2d_bind_group(
    mut commands: Commands,
    mesh2d_pipeline: Res<Mesh2dPipeline>,
    render_device: Res<RenderDevice>,
    mesh2d_uniforms: Res<BatchedInstanceBuffer<Mesh2dUniform>>,
) {
    if let Some(binding) = mesh2d_uniforms.instance_data_binding() {
        commands.insert_resource(Mesh2dBindGroup {
            value: render_device.create_bind_group(
                "mesh2d_bind_group",
                &mesh2d_pipeline.mesh_layout,
                &BindGroupEntries::single(binding),
            ),
        });
    }
}

#[derive(Component)]
pub struct Mesh2dViewBindGroup {
    pub value: BindGroup,
}

#[allow(clippy::too_many_arguments)]
pub fn prepare_mesh2d_view_bind_groups(
    mut commands: Commands,
    render_device: Res<RenderDevice>,
    mesh2d_pipeline: Res<Mesh2dPipeline>,
    view_uniforms: Res<ViewUniforms>,
    views: Query<(Entity, &Tonemapping), (With<ExtractedView>, With<Camera2d>)>,
    globals_buffer: Res<GlobalsBuffer>,
    tonemapping_luts: Res<TonemappingLuts>,
    images: Res<RenderAssets<GpuImage>>,
    fallback_image: Res<FallbackImage>,
) {
    let (Some(view_binding), Some(globals)) = (
        view_uniforms.uniforms.binding(),
        globals_buffer.buffer.binding(),
    ) else {
        return;
    };

    for (entity, tonemapping) in &views {
        let lut_bindings =
            get_lut_bindings(&images, &tonemapping_luts, tonemapping, &fallback_image);
        let view_bind_group = render_device.create_bind_group(
            "mesh2d_view_bind_group",
            &mesh2d_pipeline.view_layout,
            &BindGroupEntries::with_indices((
                (0, view_binding.clone()),
                (1, globals.clone()),
                (2, lut_bindings.0),
                (3, lut_bindings.1),
            )),
        );

        commands.entity(entity).insert(Mesh2dViewBindGroup {
            value: view_bind_group,
        });
    }
}

pub struct SetMesh2dViewBindGroup<const I: usize>;
impl<P: PhaseItem, const I: usize> RenderCommand<P> for SetMesh2dViewBindGroup<I> {
    type Param = ();
    type ViewQuery = (Read<ViewUniformOffset>, Read<Mesh2dViewBindGroup>);
    type ItemQuery = ();

    #[inline]
    fn render<'w>(
        _item: &P,
        (view_uniform, mesh2d_view_bind_group): ROQueryItem<'w, Self::ViewQuery>,
        _view: Option<()>,
        _param: SystemParamItem<'w, '_, Self::Param>,
        pass: &mut TrackedRenderPass<'w>,
    ) -> RenderCommandResult {
        pass.set_bind_group(I, &mesh2d_view_bind_group.value, &[view_uniform.offset]);

        RenderCommandResult::Success
    }
}

pub struct SetMesh2dBindGroup<const I: usize>;
impl<P: PhaseItem, const I: usize> RenderCommand<P> for SetMesh2dBindGroup<I> {
    type Param = SRes<Mesh2dBindGroup>;
    type ViewQuery = ();
    type ItemQuery = ();

    #[inline]
    fn render<'w>(
        item: &P,
        _view: (),
        _item_query: Option<()>,
        mesh2d_bind_group: SystemParamItem<'w, '_, Self::Param>,
        pass: &mut TrackedRenderPass<'w>,
    ) -> RenderCommandResult {
        let mut dynamic_offsets: [u32; 1] = Default::default();
        let mut offset_count = 0;
        if let Some(dynamic_offset) = item.extra_index().as_dynamic_offset() {
            dynamic_offsets[offset_count] = dynamic_offset.get();
            offset_count += 1;
        }
        pass.set_bind_group(
            I,
            &mesh2d_bind_group.into_inner().value,
            &dynamic_offsets[..offset_count],
        );
        RenderCommandResult::Success
    }
}

pub struct DrawMesh2d;
impl<P: PhaseItem> RenderCommand<P> for DrawMesh2d {
    type Param = (
        SRes<RenderAssets<RenderMesh>>,
        SRes<RenderMesh2dInstances>,
        SRes<MeshAllocator>,
    );
    type ViewQuery = ();
    type ItemQuery = ();

    #[inline]
    fn render<'w>(
        item: &P,
        _view: (),
        _item_query: Option<()>,
        (meshes, render_mesh2d_instances, mesh_allocator): SystemParamItem<'w, '_, Self::Param>,
        pass: &mut TrackedRenderPass<'w>,
    ) -> RenderCommandResult {
        let meshes = meshes.into_inner();
        let render_mesh2d_instances = render_mesh2d_instances.into_inner();
        let mesh_allocator = mesh_allocator.into_inner();

        let Some(RenderMesh2dInstance { mesh_asset_id, .. }) =
            render_mesh2d_instances.get(&item.entity())
        else {
            return RenderCommandResult::Skip;
        };
        let Some(gpu_mesh) = meshes.get(*mesh_asset_id) else {
            return RenderCommandResult::Skip;
        };
        let Some(vertex_buffer_slice) = mesh_allocator.mesh_vertex_slice(mesh_asset_id) else {
            return RenderCommandResult::Skip;
        };

        pass.set_vertex_buffer(0, vertex_buffer_slice.buffer.slice(..));

        let batch_range = item.batch_range();
        match &gpu_mesh.buffer_info {
            RenderMeshBufferInfo::Indexed {
                index_format,
                count,
            } => {
                let Some(index_buffer_slice) = mesh_allocator.mesh_index_slice(mesh_asset_id)
                else {
                    return RenderCommandResult::Skip;
                };

                pass.set_index_buffer(index_buffer_slice.buffer.slice(..), 0, *index_format);

                pass.draw_indexed(
                    index_buffer_slice.range.start..(index_buffer_slice.range.start + count),
                    vertex_buffer_slice.range.start as i32,
                    batch_range.clone(),
                );
            }
            RenderMeshBufferInfo::NonIndexed => {
                pass.draw(0..gpu_mesh.vertex_count, batch_range.clone());
            }
        }
        RenderCommandResult::Success
    }
}