feature: hud (#37)

* build: replace FetchContent with direct includes for glm

* build(deps): replace SOIL2 with stb for image and vorbis decoding

* build: replace FetchContent with direct includes for toml++

* refactor(texture): bump block texture size to 512 and disable items tab

* feat: add hotbar system with ItemStack and RowLayout

* fix(ui): restrict Ctrl key handling to typing mode

* feat(ui): add widget border support and highlight selected hotbar slot

* fix(ui): update borders on scale change and correct Image width

- Added update_border() calls in set_scale() for Button, ChatBox, Image, Label, and TextField.
- Fixed Image::width() to multiply by width instead of height.
- Added early return in Widget::update_border() if border is not supported.

* feat(ui): add border visual feedback on slider drag

* feat(ui): add ItemSlot widget and refactor hotbar to use it

* feat(world_scene): add inventory UI and pause type enum

* feat(ui): add item slot tooltip and make window size static

* fix(ui label): account for background offset in width/height

* refactor(ui): centralize item info label in inventory and add update

* feat(ui): add hotbar interaction in inventory UI

* fix(inventory-ui): correct row creation logic for first item

The loop starting at index 0 created a new row on the first iteration
(i % 10 == 0). Shift the loop to start at 1 and adjust modulus condition
to create rows correctly every 10 items.

* feat(block): add name_key for localized block names

Add `name_key` field to all block TOML definitions, enabling per-block localization. Update `BlockData` struct and `BlockManager` with `local_name()` method that returns a localized string via the translation system. Add corresponding entries to `en_US.json` and `zh_CN.json`. Modify inventory UI to use the localized name when displaying block info, and fix a bug where emptying a hotbar slot did not clear the item locally.

* fix(ui): change inventory item slot color from white to gray

* fix(ui): add missing include for std::array

include/glm/ext/quaternion_exponential.inl

@@ -0,0 +1,89 @@
+#include "scalar_constants.hpp"
+
+namespace glm
+{
+	template<typename T, qualifier Q>
+	GLM_FUNC_QUALIFIER qua<T, Q> exp(qua<T, Q> const& q)
+	{
+		vec<3, T, Q> u(q.x, q.y, q.z);
+		T const Angle = glm::length(u);
+		if (Angle < epsilon<T>())
+			return qua<T, Q>();
+
+		vec<3, T, Q> const v(u / Angle);
+		return qua<T, Q>(cos(Angle), sin(Angle) * v);
+	}
+
+	template<typename T, qualifier Q>
+	GLM_FUNC_QUALIFIER qua<T, Q> log(qua<T, Q> const& q)
+	{
+		vec<3, T, Q> u(q.x, q.y, q.z);
+		T Vec3Len = length(u);
+
+		if (Vec3Len < epsilon<T>())
+		{
+			if(q.w > static_cast<T>(0))
+				return qua<T, Q>::wxyz(log(q.w), static_cast<T>(0), static_cast<T>(0), static_cast<T>(0));
+			else if(q.w < static_cast<T>(0))
+				return qua<T, Q>::wxyz(log(-q.w), pi<T>(), static_cast<T>(0), static_cast<T>(0));
+			else
+				return qua<T, Q>::wxyz(std::numeric_limits<T>::infinity(), std::numeric_limits<T>::infinity(), std::numeric_limits<T>::infinity(), std::numeric_limits<T>::infinity());
+		}
+		else
+		{
+			T t = atan(Vec3Len, T(q.w)) / Vec3Len;
+			T QuatLen2 = Vec3Len * Vec3Len + q.w * q.w;
+			return qua<T, Q>::wxyz(static_cast<T>(0.5) * log(QuatLen2), t * q.x, t * q.y, t * q.z);
+		}
+	}
+
+	template<typename T, qualifier Q>
+	GLM_FUNC_QUALIFIER qua<T, Q> pow(qua<T, Q> const& x, T y)
+	{
+		//Raising to the power of 0 should yield 1
+		//Needed to prevent a division by 0 error later on
+		if(y > -epsilon<T>() && y < epsilon<T>())
+			return qua<T, Q>::wxyz(1,0,0,0);
+
+		//To deal with non-unit quaternions
+		T magnitude = sqrt(x.x * x.x + x.y * x.y + x.z * x.z + x.w *x.w);
+
+		T Angle;
+		if(abs(x.w / magnitude) > cos_one_over_two<T>())
+		{
+			//Scalar component is close to 1; using it to recover angle would lose precision
+			//Instead, we use the non-scalar components since sin() is accurate around 0
+
+			//Prevent a division by 0 error later on
+			T VectorMagnitude = x.x * x.x + x.y * x.y + x.z * x.z;
+			//Despite the compiler might say, we actually want to compare
+			//VectorMagnitude to 0. here; we could use denorm_int() compiling a
+			//project with unsafe maths optimizations might make the comparison
+			//always false, even when VectorMagnitude is 0.
+			if (VectorMagnitude < (std::numeric_limits<T>::min)()) {
+				//Equivalent to raising a real number to a power
+				return qua<T, Q>::wxyz(pow(x.w, y), 0, 0, 0);
+			}
+
+			Angle = asin(sqrt(VectorMagnitude) / magnitude);
+		}
+		else
+		{
+			//Scalar component is small, shouldn't cause loss of precision
+			Angle = acos(x.w / magnitude);
+		}
+
+		T NewAngle = Angle * y;
+		T Div = sin(NewAngle) / sin(Angle);
+		T Mag = pow(magnitude, y - static_cast<T>(1));
+		return qua<T, Q>::wxyz(cos(NewAngle) * magnitude * Mag, x.x * Div * Mag, x.y * Div * Mag, x.z * Div * Mag);
+	}
+
+	template<typename T, qualifier Q>
+	GLM_FUNC_QUALIFIER qua<T, Q> sqrt(qua<T, Q> const& x)
+	{
+		return pow(x, static_cast<T>(0.5));
+	}
+}//namespace glm
+
+