/* sdl_gpu_test.c -- -lSDL3 -lm */

#define SDL_MAIN_USE_CALLBACKS 1
#include <SDL3/SDL_main.h>
#include <SDL3/SDL_gpu.h>
#include <math.h>
#include <stdio.h>
#include <err.h>

SDL_Window *win;
SDL_GPUDevice *gpudev;
SDL_GPUGraphicsPipeline *pipeline_fill;
SDL_GPUBuffer *buf_vert;
SDL_GPUTransferBuffer *buf_vert_xfer;

typedef struct {
	float x, y;
	float u, v;
} Vertex;

Vertex vertices[] = {
	{ -1, -1,  0, 0, },
	{ 0, -1,  1, 0, },
	{ 0, 0,  1, 1, },

	{ -1, -1,  0, 0, },
	{ -1, 0,  0, 1, },
	{ 0, 0,  1, 1, },

	{ 0, 0,  0, 0, },
	{ 1, 0,  1, 0, },
	{ 1, 1,  1, 1, },

	{ 0, 0,  0, 0, },
	{ 0, 1,  0, 1, },
	{ 1, 1,  1, 1, },

};

typedef struct {
	SDL_GPUShaderStage stage;
	Uint32 n_sampler, n_uniformbuf, n_storagebuf, n_storagetex;
} ShaderDesc;

SDL_GPUShader* load_shader(SDL_GPUDevice* dev, const char *path, ShaderDesc *desc) {
	size_t code_size;
	void *code = SDL_LoadFile(path, &code_size);
	if (!code) {
		fprintf(stderr, "couldn't read file: %s\n", SDL_GetError());
		return NULL;
	}

	SDL_GPUShader *sh = SDL_CreateGPUShader(dev, &(SDL_GPUShaderCreateInfo) {
		.code = code,
		.code_size = code_size,
		.entrypoint = "main",
		.format = SDL_GPU_SHADERFORMAT_SPIRV,
		.stage = desc->stage,
		.num_samplers = desc->n_sampler,
		.num_uniform_buffers = desc->n_uniformbuf,
		.num_storage_buffers = desc->n_storagebuf,
		.num_storage_textures = desc->n_storagetex
	});

	if (!sh) return NULL;
	SDL_free(code);
	return sh;
}

#define CHECK(t, msg) do { if (!(t)) { fprintf(stderr, "%s: %s\n", msg, SDL_GetError()); return SDL_APP_FAILURE; } } while(0)

SDL_AppResult SDL_AppInit(void **appstate, int argc, char **argv) {
	(void)argc;
	(void)argv;
	(void)appstate;
	/* TODO: support more than just SPIR-V */
	win = SDL_CreateWindow("Hello, world!", 1024, 768, SDL_WINDOW_RESIZABLE | SDL_WINDOW_VULKAN);
	CHECK(win, "failed to create window");
	gpudev = SDL_CreateGPUDevice(SDL_GPU_SHADERFORMAT_SPIRV, 0, NULL);
	CHECK(gpudev, "failed to create gpu device");
	CHECK(SDL_ClaimWindowForGPUDevice(gpudev, win), "failed to claim window");

	SDL_GPUShader *shader_vert = load_shader(gpudev, "shader/tri.vert.spv", &(ShaderDesc) {
			.stage = SDL_GPU_SHADERSTAGE_VERTEX,
			.n_uniformbuf = 1
		});
	CHECK(shader_vert, "load vertex shader");
	SDL_GPUShader *shader_frag = load_shader(gpudev, "shader/tri.frag.spv", &(ShaderDesc) { SDL_GPU_SHADERSTAGE_FRAGMENT });
	CHECK(shader_frag, "load fragment shader");

	buf_vert = SDL_CreateGPUBuffer(gpudev, &(SDL_GPUBufferCreateInfo) {
		.usage = SDL_GPU_BUFFERUSAGE_VERTEX,
		.size = sizeof(vertices),
		.props = 0
	});
	CHECK(buf_vert, "vertex buffer");

	buf_vert_xfer = SDL_CreateGPUTransferBuffer(
		gpudev,
		&(SDL_GPUTransferBufferCreateInfo) {
			.usage = SDL_GPU_TRANSFERBUFFERUSAGE_UPLOAD,
			.size = sizeof(vertices),
			.props = 0
		}
	);

	CHECK(buf_vert_xfer, "vertex transfer buffer");

	void *map = SDL_MapGPUTransferBuffer(gpudev, buf_vert_xfer, false);
	SDL_memcpy(map, vertices, sizeof(vertices));
	SDL_UnmapGPUTransferBuffer(gpudev, buf_vert_xfer);

	SDL_GPUCommandBuffer *cmdbuf = SDL_AcquireGPUCommandBuffer(gpudev);
	CHECK(cmdbuf, "gpu command buffer acquisition");
	SDL_GPUCopyPass *copypass = SDL_BeginGPUCopyPass(cmdbuf);
	CHECK(copypass, "gpu copy pass");

	SDL_UploadToGPUBuffer(
		copypass,
		&(SDL_GPUTransferBufferLocation) {
			.transfer_buffer = buf_vert_xfer,
			.offset = 0
		},
		&(SDL_GPUBufferRegion) {
			.buffer = buf_vert,
			.offset = 0,
			.size = sizeof(vertices)
		},
		false
	);

	SDL_EndGPUCopyPass(copypass);
	SDL_SubmitGPUCommandBuffer(cmdbuf);
	SDL_ReleaseGPUTransferBuffer(gpudev, buf_vert_xfer);

	pipeline_fill = SDL_CreateGPUGraphicsPipeline(gpudev, &(SDL_GPUGraphicsPipelineCreateInfo) {
		.target_info = {
			.num_color_targets = 1,
			.color_target_descriptions = (SDL_GPUColorTargetDescription[]) {{
				.format = SDL_GetGPUSwapchainTextureFormat(gpudev, win),
			       .blend_state = {
				       .enable_blend = true,
				       .color_blend_op = SDL_GPU_BLENDOP_ADD,
				       .alpha_blend_op = SDL_GPU_BLENDOP_ADD,
				       .src_color_blendfactor = SDL_GPU_BLENDFACTOR_SRC_ALPHA,
				       .dst_color_blendfactor = SDL_GPU_BLENDFACTOR_ONE_MINUS_SRC_ALPHA,
				       .src_alpha_blendfactor = SDL_GPU_BLENDFACTOR_SRC_ALPHA,
				       .dst_alpha_blendfactor = SDL_GPU_BLENDFACTOR_ONE_MINUS_SRC_ALPHA,
			       }
			}},
		},
		.primitive_type = SDL_GPU_PRIMITIVETYPE_TRIANGLELIST,
		.vertex_shader = shader_vert,
		.fragment_shader = shader_frag,
		.vertex_input_state = {
			.num_vertex_buffers = 1,
			.vertex_buffer_descriptions = &(SDL_GPUVertexBufferDescription) {
				.slot = 0,
				.pitch = sizeof(Vertex),
				.input_rate = SDL_GPU_VERTEXINPUTRATE_VERTEX,
				.instance_step_rate = 0,
			},
			.num_vertex_attributes = 2,
			.vertex_attributes = (SDL_GPUVertexAttribute[]) {
				{
					.location = 0,
					.buffer_slot = 0,
					.format = SDL_GPU_VERTEXELEMENTFORMAT_FLOAT2,
					.offset = 0
				},
				{
					.location = 1,
					.buffer_slot = 0,
					.format = SDL_GPU_VERTEXELEMENTFORMAT_FLOAT2,
					.offset = sizeof(float) * 2
				}
			}
		},
		.rasterizer_state = {
			.fill_mode = SDL_GPU_FILLMODE_FILL
		},
		.props = 0
	});

	CHECK(pipeline_fill, "graphics pipeline");

	SDL_ReleaseGPUShader(gpudev, shader_frag);
	SDL_ReleaseGPUShader(gpudev, shader_vert);

	return SDL_APP_CONTINUE;
}

SDL_AppResult SDL_AppEvent(void *appstate, SDL_Event *event) {
	(void)appstate;
	switch (event->type) {
	case SDL_EVENT_QUIT:
		puts("Quit!");
		return SDL_APP_SUCCESS;
	default:
		break;
	}
	return SDL_APP_CONTINUE;
}

SDL_AppResult SDL_AppIterate(void *appstate) {
	(void)appstate;

	SDL_GPUCommandBuffer *cmdbuf = SDL_AcquireGPUCommandBuffer(gpudev);
	CHECK(cmdbuf, "gpu cmd buffer acquisition");

	SDL_GPUTexture *swapchain;
	Uint32 swc_width, swc_height;
	CHECK(SDL_WaitAndAcquireGPUSwapchainTexture(cmdbuf, win, &swapchain, &swc_width, &swc_height), "swapchain texture acquisition");
	CHECK(swapchain, "nil swapchain texture");

	float how = (float)swc_height / (float)swc_width;
	float woh = (float)swc_width / (float)swc_height;
	float scale = (32 * 2) / fminf(swc_width, swc_height);
	float scale_x = (swc_height < swc_width ? how : 1) * scale;
	float scale_y = (swc_height < swc_width ? 1 : woh) * scale;
	float cam_mat[4][4] = {
		{ scale_x, 0, 0, 0, },
		{ 0, scale_y, 0, 0, },
		{ 0, 0, 1, 0, },
		{ 0, 0, 0, 1, },
	};
	SDL_PushGPUVertexUniformData(cmdbuf, 0, &cam_mat, sizeof(cam_mat));
	SDL_GPURenderPass *render_pass = SDL_BeginGPURenderPass(
		cmdbuf,
		&(SDL_GPUColorTargetInfo) {
			.texture = swapchain,
			.clear_color = { 0, 0, 0, 1 },
			.load_op = SDL_GPU_LOADOP_CLEAR,
			.store_op = SDL_GPU_STOREOP_STORE,
		},
		1,
		NULL
	);
        SDL_BindGPUGraphicsPipeline(render_pass, pipeline_fill);
	SDL_BindGPUVertexBuffers(render_pass, 0, &(SDL_GPUBufferBinding) { .buffer = buf_vert, .offset = 0 }, 1);
        SDL_DrawGPUPrimitives(render_pass, sizeof(vertices) / sizeof(Vertex), 1, 0, 0);
        SDL_EndGPURenderPass(render_pass);

	SDL_SubmitGPUCommandBuffer(cmdbuf);
	return SDL_APP_CONTINUE;
}

void SDL_AppQuit(void *appstate, SDL_AppResult result) {
	(void)appstate;
	(void)result;
	SDL_ReleaseGPUGraphicsPipeline(gpudev, pipeline_fill);
	SDL_ReleaseWindowFromGPUDevice(gpudev, win);
	SDL_DestroyGPUDevice(gpudev);
	SDL_DestroyWindow(win);
}