#include "data/font_info.h"
INCBIN(_font_texture, "data/font.ugg");
struct ugg *font_texture_data = (struct ugg*)_font_texture_data;

/* Overlay state structure */
struct overlay {
	GLuint vao;
	GLuint vbo;
	GLuint ebo;
	GLuint program;
	GLint  loc_proj;
	GLint  loc_tex;
	GLint  loc_color;
	GLint  loc_pos_offset;
	GLuint font_texture;
	GLuint white_texture;
};

static struct overlay overlay_state;
static float mat[16];

/* ------------------------------------------------------------------------- */
/* Shader sources for a top-left orthographic approach                      */
/* ------------------------------------------------------------------------- */
static const char* overlay_vertex_shader_src =
"#version 140\n"
"in vec2 in_pos;\n"
"in vec2 in_uv;\n"
"uniform mat4 u_projection;\n"
"uniform vec2 u_pos_offset;\n"
"out vec2 v_uv;\n"
"void main() {\n"
"    vec2 pos = in_pos + u_pos_offset;\n"
"    gl_Position = u_projection * vec4(pos, 0.0, 1.0);\n"
"    v_uv = in_uv;\n"
"}\n";

static const char* overlay_fragment_shader_src =
"#version 140\n"
"uniform sampler2D u_font_texture;\n"
"uniform vec4 u_color;\n"
"in vec2 v_uv;\n"
"out vec4 frag_color;\n"
"void main() {\n"
"	vec4 tex_sample = texture(u_font_texture, v_uv);\n"
"	float alpha = tex_sample.a;\n"
"	frag_color = vec4(u_color.rgb * alpha, u_color.a * alpha);\n"
"}\n";

/* ------------------------------------------------------------------------- */
/* Shader helpers                                                            */
/* ------------------------------------------------------------------------- */
__attribute__((cold, noinline, section(".init_section")))
static GLuint overlay_compile_shader(const char *source, GLenum type) {
	GLuint shader = glCreateShader(type);
	if(shader == 0) {
		fprintf(stderr, "Error creating shader of type %d.\n", type);
		return 0;
	}

	glShaderSource(shader, 1, &source, 0);
	glCompileShader(shader);

	/* Check for compilation errors */
	GLint success;
	glGetShaderiv(shader, GL_COMPILE_STATUS, &success);
	if(!success) {
		GLint log_size = 0;
		glGetShaderiv(shader, GL_INFO_LOG_LENGTH, &log_size);
		char log[2048];
		glGetShaderInfoLog(shader, log_size, 0, log);
		fprintf(stderr, "Shader compilation failed:\n%s\n", log);
		glDeleteShader(shader);
		return 0;
	}

	return shader;
}

__attribute__((cold, noinline, section(".init_section")))
static GLuint overlay_create_program(const char *vs_source, const char *fs_source) {
	GLuint vs = overlay_compile_shader(vs_source, GL_VERTEX_SHADER);
	GLuint fs = overlay_compile_shader(fs_source, GL_FRAGMENT_SHADER);
	GLuint prog = glCreateProgram();

	glAttachShader(prog, vs);
	glAttachShader(prog, fs);
	glLinkProgram(prog);

	/* Check for linking errors */
	GLint success;
	glGetProgramiv(prog, GL_LINK_STATUS, &success);
	if(!success) {
		/* Get and print log */
		GLint log_size = 0;
		glGetProgramiv(prog, GL_INFO_LOG_LENGTH, &log_size);
		char log[2048];
		glGetProgramInfoLog(prog, log_size, 0, log);
		fprintf(stderr, "Program linking failed:\n%s\n", log);
		glDeleteProgram(prog);
		glDeleteShader(vs);
		glDeleteShader(fs);
		return 0;
	}

	/* Shaders can be deleted after linking */
	glDeleteShader(vs);
	glDeleteShader(fs);

	return prog;
}

static void overlay_make_ortho_top_left(float w, float h, float m[16]) {
	m[0] =  2.0f / w;
	m[5] = -2.0f / h;
	m[10] = -1.0f;
	m[12] = -1.0f;
	m[13] =  1.0f;
	m[15] =  1.0f;
}

#define MAX_GLYPHS 128  // NOTE(peter): max glyphs per ROW
#define MAX_VERTICES (MAX_PROFILING_ENTRIES * MAX_GLYPHS * 4 * 4)
float vertices[MAX_VERTICES] __attribute__((section(".bss")));

static void overlay_init(void) {
	overlay_state.program			= overlay_create_program(overlay_vertex_shader_src, overlay_fragment_shader_src);
	overlay_state.loc_proj			= glGetUniformLocation(overlay_state.program, "u_projection");
	overlay_state.loc_tex			= glGetUniformLocation(overlay_state.program, "u_font_texture");
	overlay_state.loc_color			= glGetUniformLocation(overlay_state.program, "u_color");
	overlay_state.loc_pos_offset	= glGetUniformLocation(overlay_state.program, "u_pos_offset");

	glGenVertexArrays(1, &overlay_state.vao);
	glBindVertexArray(overlay_state.vao);

	glGenBuffers(1, &overlay_state.vbo);
	glBindBuffer(GL_ARRAY_BUFFER, overlay_state.vbo);
	glBufferData(GL_ARRAY_BUFFER, MAX_VERTICES * sizeof(float), 0, GL_DYNAMIC_DRAW);

	glGenBuffers(1, &overlay_state.ebo);

	unsigned short indices[MAX_PROFILING_ENTRIES * MAX_GLYPHS * 6];
	size_t index_offset = 0;
	size_t vertex_offset = 0;

	for(size_t i = 0; i < MAX_PROFILING_ENTRIES * MAX_GLYPHS; i++) {
		indices[index_offset++] = vertex_offset + 0;
		indices[index_offset++] = vertex_offset + 1;
		indices[index_offset++] = vertex_offset + 2;
		indices[index_offset++] = vertex_offset + 0;
		indices[index_offset++] = vertex_offset + 2;
		indices[index_offset++] = vertex_offset + 3;
		vertex_offset += 4;
	}

	glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, overlay_state.ebo);
	glBufferData(GL_ELEMENT_ARRAY_BUFFER, sizeof(indices), indices, GL_STATIC_DRAW);

	GLint in_pos_attrib = glGetAttribLocation(overlay_state.program, "in_pos");
	GLint in_uv_attrib  = glGetAttribLocation(overlay_state.program, "in_uv");

	glEnableVertexAttribArray(in_pos_attrib);
	glVertexAttribPointer(in_pos_attrib, 2, GL_FLOAT, GL_FALSE, 4 * sizeof(float), (void*)0);
	glEnableVertexAttribArray(in_uv_attrib);
	glVertexAttribPointer(in_uv_attrib, 2, GL_FLOAT, GL_FALSE, 4 * sizeof(float), (void*)(2 * sizeof(float)));

	glBindVertexArray(0);
	glBindBuffer(GL_ARRAY_BUFFER, 0);
	glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, 0);


	uint8_t *rgba_data = mks_alloc(512 * 512 * 4);
	uint8_t *dst = rgba_data;
	for(uint32_t i = 0; i < 512 * 512; ++i) {
		uint8_t alpha = font_texture_data->data[i];
		if(alpha) {
			*dst++ = 255;
			*dst++ = 255;
			*dst++ = 255;
			*dst++ = alpha;

		} else {
			*dst++ = 0;
			*dst++ = 0;
			*dst++ = 0;
			*dst++ = 0;
		}
	}
	glGenTextures(1, &overlay_state.font_texture);
	glBindTexture(GL_TEXTURE_2D, overlay_state.font_texture);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
	// glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR_MIPMAP_LINEAR);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
	// glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
	// glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
	glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, 512, 512, 0, GL_RGBA, GL_UNSIGNED_BYTE, rgba_data);
	glGenerateMipmap(GL_TEXTURE_2D);
	mks_free(rgba_data);

	glGenTextures(1, &overlay_state.white_texture);
	glBindTexture(GL_TEXTURE_2D, overlay_state.white_texture);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
	glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
	uint32_t white[] = { 0xffffffff };
	glTexImage2D(GL_TEXTURE_2D, 0, GL_SRGB8_ALPHA8, 1, 1, 0, GL_RGBA, GL_UNSIGNED_INT_8_8_8_8, white);

	glBindTexture(GL_TEXTURE_2D, 0);
}

__attribute__((cold, noinline, section(".init_section")))
static void overlay_shutdown(void) {
	glDeleteProgram(overlay_state.program);
	glDeleteBuffers(1, &overlay_state.vbo);
	glDeleteBuffers(1, &overlay_state.ebo);
	glDeleteVertexArrays(1, &overlay_state.vao);
	glDeleteTextures(1, &overlay_state.font_texture);
}

static void overlay_render_rect(float x1, float y1, float x2, float y2, float r, float g, float b, float a) {
	float vertices[4 * 4] = {
		/* pos.x, pos.y, u, v */
		x1, y1,  0.f, 0.f, // Top-left
		x2, y1,  1.f, 0.f, // Top-right
		x2, y2,  1.f, 1.f, // Bottom-right
		x1, y2,  0.f, 1.f  // Bottom-left
	};

	glBindVertexArray(overlay_state.vao);
	glBindBuffer(GL_ARRAY_BUFFER, overlay_state.vbo);
	glBufferSubData(GL_ARRAY_BUFFER, 0, sizeof(vertices), vertices);
	glUseProgram(overlay_state.program);
	glUniform4f(overlay_state.loc_color, r, g, b, a);
	glActiveTexture(GL_TEXTURE0);
	glBindTexture(GL_TEXTURE_2D, overlay_state.white_texture);
	glUniform1i(overlay_state.loc_tex, 0);
	glUniform2f(overlay_state.loc_pos_offset, 0.0f, 0.0f);
	glDrawElements(GL_TRIANGLES, 6, GL_UNSIGNED_SHORT, 0);
	glBindTexture(GL_TEXTURE_2D, 0);
	glBindVertexArray(0);
}

__attribute__((always_inline))
static inline void overlay_render_text_line(uint8_t **lines, float x, float y, float r, float g, float b, float a) {
	uint32_t vertex_offset = 0;
	uint32_t glyph_count = 0;

	glBindVertexArray(overlay_state.vao);
	glBindBuffer(GL_ARRAY_BUFFER, overlay_state.vbo);

	float target_width = 24.f * .40f;
	float target_height = 48.f * .40f;

	// Convert font units to pixels
	float _scale = 48.0f / (1900 - (-480));
	float scaled_ascent = 1900 * _scale;
	float scale_x = target_width / 24.0f;
	float scale_y = target_height / 48.0f;

	// **Apply baseline correction once (not per character!)**
	y += scaled_ascent * scale_y;

	for(int i = 0; i < MAX_PROFILING_ENTRIES; i++) {
		uint8_t *text = lines[i];
		if(!text) continue;

		float cx = x;
		while(*text) {
			uint8_t c = *text++;
			struct glyph_info *g = &glyph_data[c];

			// Normalize texture coordinates
			float u0 = g->x / 512.f;
			float v0 = g->y / 512.f;
			float u1 = (g->x + g->width) / 512.f;
			float v1 = (g->y + g->height) / 512.f;

			// Apply width & height scaling
			float glyph_width = g->width * scale_x;
			float glyph_height = g->height * scale_y;

			// **Fix baseline positioning**
			float x0 = roundf(cx + (g->x_offset * scale_x));
			float y0 = roundf(y + (g->y_offset * scale_y));
			float x1 = x0 + glyph_width;
			float y1 = y0 + glyph_height;

			// Generate quad for the character
			vertices[vertex_offset + 0] = x0;
			vertices[vertex_offset + 1] = y0;
			vertices[vertex_offset + 2] = u0;
			vertices[vertex_offset + 3] = v0;

			vertices[vertex_offset + 4] = x1;
			vertices[vertex_offset + 5] = y0;
			vertices[vertex_offset + 6] = u1;
			vertices[vertex_offset + 7] = v0;

			vertices[vertex_offset + 8] = x1;
			vertices[vertex_offset + 9] = y1;
			vertices[vertex_offset + 10] = u1;
			vertices[vertex_offset + 11] = v1;

			vertices[vertex_offset + 12] = x0;
			vertices[vertex_offset + 13] = y1;
			vertices[vertex_offset + 14] = u0;
			vertices[vertex_offset + 15] = v1;

			// Move cursor forward
			cx += g->advance * scale_x;
			vertex_offset += 16;
			glyph_count++;
		}
		y += target_height;  // Move to next row
	}


	if(glyph_count > 0) {
		glBindVertexArray(overlay_state.vao);
		glBindBuffer(GL_ARRAY_BUFFER, overlay_state.vbo);
		glBufferSubData(GL_ARRAY_BUFFER, 0, vertex_offset * sizeof(float), vertices);

		glActiveTexture(GL_TEXTURE0);
		glBindTexture(GL_TEXTURE_2D, overlay_state.font_texture);
		glUniform1i(overlay_state.loc_tex, 0);
		glUniform4f(overlay_state.loc_color, r, g, b, a);
		glDrawElements(GL_TRIANGLES, glyph_count * 6, GL_UNSIGNED_SHORT, 0);
		glBindVertexArray(0);
	}
}

static void overlay_render(float rect_x, float rect_y, float rect_w, float rect_h, uint8_t **lines, int window_w, int window_h) {
	overlay_make_ortho_top_left((float)window_w, (float)window_h, mat);
	glViewport(0.f, 0.f, window_w, window_h);
	glUseProgram(overlay_state.program);
	glUniformMatrix4fv(overlay_state.loc_proj, 1, GL_FALSE, mat);

	overlay_render_rect(rect_x, rect_y, rect_x + rect_w, rect_y + rect_h, .01f, .013f, .04f, .6f);
	overlay_render_text_line(lines, rect_x + 2.f, rect_y + 2.f, 0.f, 0.f, 0.f, 1.f);
	overlay_render_text_line(lines, rect_x, rect_y, 1.f, 1.f, 1.f, 1.f);
}

/* [=]===^=[ debug_render ]=================================================================^===[=] */
__attribute__((section(".bss")))
uint8_t *debug_lines[MAX_PROFILING_ENTRIES];
static void debug_render(void) {
	if(state.overlay) {
		size_t offset = 0;
		for(uint32_t i = 0; i < MAX_PROFILING_ENTRIES; ++i) {
			if(state.debug.timings[i].count) {
				debug_lines[i] = &debug_line_buffer[offset];
				offset += stbsp_sprintf((char *)&debug_line_buffer[offset], "%25s: cycles=%7" PRIu64 ", count=%2u, cycles/count=%7" PRIu64, state.debug.timings[i].name, state.debug.timings[i].cycles, state.debug.timings[i].count, state.debug.timings[i].cycles / state.debug.timings[i].count) + 1;
			} else {
				debug_lines[i] = 0;
			}
		}
		overlay_render(20.f, 20.f, 705.f, 280.f, debug_lines, state.screen_width, state.screen_height);

		offset = 0;
		for(uint32_t i = 0; i < MAX_PROFILING_ENTRIES; ++i) {
			debug_lines[i] = 0;
		}
		debug_lines[0] = &debug_line_buffer[offset];
		offset += stbsp_sprintf((char *)&debug_line_buffer[offset], "Total memory allocated: %" PRIu64, state.total_allocated) + 1;
		overlay_render(20.f, 340.f, 500.f, 200.f, debug_lines, state.screen_width, state.screen_height);
	}
}

/* [=]===^=[ reset_profiling_data ]=================================================================^===[=] */
static inline void reset_profiling_data(void) {
	memset(&state.debug, 0, sizeof(state.debug));
}