port 2V2-d.2: Granny runtime on libgr2, main character model animates

Implement the Granny SDK 2.11 API the client uses on top of libgr2
(platform/EterGrnLib/GrannyRuntime.cpp, shim granny.h declared from the
SDK header): file sections, type-definition vertex conversion, material
textures by Usage, mesh/bone bindings, deformer, control clock/loop/ease
and weighted SRT sampling. Port EterGrnLib (LODController, Material, Mesh,
Model, ModelInstance*, Motion, Thing, ThingInstance, Util) plus EterLib
GrpObjectInstance/GrpCollisionObject/CollisionData verbatim and drop their
platform skeletons and pending stand-ins; register the gr2 resource
factory; vertex/index buffers are CPU-side until the 2V2-e render adapter.

app.UpdateGame/RenderGame now run 40250's bodies, reaching the
CPythonApplication members through their singletons (the platform has no
app object); CPythonGraphic joins GameSingletons in member order, and
GetMousePosition reads the host cursor.

port.login_flow asserts the main instance's PART_MAIN model has meshes,
its diffuse image loaded and the WAIT motion moves the bones between
frames; offline and live pass, port_gate macos PASS.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

libgr2/src/gr2_skeleton.cpp

@@ -57,32 +57,43 @@ void m3_mul(const float a[9], const float b[9], float r[9]) {
 } // namespace
 
 // granny_transform (68B: u32 Flags; f32 Position[3]; f32 Orientation[4]; f32 ScaleShear[3][3])
-// → 行主序 4x4，上 3x3 = transpose(U3)，平移在 elem 12..14。BuildCompositeTransform4x4 语义。
-void File::Impl::transform_to_composite(Ref tf, Mat4& out, uint32_t& flags) const {
+void File::Impl::read_transform(Ref tf, Transform& out) const {
+    out = Transform{};
     const uint8_t* p = at(tf, 68);
-    out = Mat4{};
-    out[0] = out[5] = out[10] = out[15] = 1.0f;
-    flags = 0;
     if (!p) return;
+    std::memcpy(&out.flags, p, 4);
+    std::memcpy(out.position, p + 4, 12);
+    std::memcpy(out.orientation, p + 16, 16);
+    std::memcpy(out.scale_shear, p + 32, 36);
+}
 
-    auto f32 = [&](size_t off) { float v; std::memcpy(&v, p + off, 4); return v; };
-    flags = rd_u32(tf);
-    float pos[3] = { f32(4), f32(8), f32(12) };
-    float ori[4] = { f32(16), f32(20), f32(24), f32(28) };
-    float ss[9];
-    for (int i = 0; i < 9; ++i) ss[i] = f32(32 + size_t(i) * 4);
-
+// → 行主序 4x4，上 3x3 = transpose(U3)，平移在 elem 12..14。BuildCompositeTransform4x4 语义。
+Mat4 compose(const Transform& t) {
     float q3[9];
-    quat_to_m3(ori, q3);
+    quat_to_m3(t.orientation, q3);
     float u3[9];
-    if (flags & 0x4 /* HasScaleShear */) m3_mul(q3, ss, u3);
+    if (t.flags & 0x4 /* HasScaleShear */) m3_mul(q3, t.scale_shear, u3);
     else std::memcpy(u3, q3, sizeof u3);
 
+    Mat4 out{};
     // C 的上 3x3 = transpose(U3)
     out[0] = u3[0]; out[1] = u3[3]; out[2] = u3[6]; out[3] = 0;
     out[4] = u3[1]; out[5] = u3[4]; out[6] = u3[7]; out[7] = 0;
     out[8] = u3[2]; out[9] = u3[5]; out[10] = u3[8]; out[11] = 0;
-    out[12] = pos[0]; out[13] = pos[1]; out[14] = pos[2]; out[15] = 1;
+    out[12] = t.position[0]; out[13] = t.position[1]; out[14] = t.position[2]; out[15] = 1;
+    return out;
+}
+
+void File::Impl::transform_to_composite(Ref tf, Mat4& out, uint32_t& flags) const {
+    Transform t;
+    read_transform(tf, t);
+    if (!at(tf, 68)) {
+        out = Mat4{1,0,0,0, 0,1,0,0, 0,0,1,0, 0,0,0,1};
+        flags = 0;
+        return;
+    }
+    out = compose(t);
+    flags = t.flags;
 }
 
 // skeleton: {String Name; RefToArray Bones(bone[]); int32 LODType; VariantRef ExtendedData}
@@ -118,9 +129,9 @@ bool File::Impl::extract_skeleton(Ref skel_obj, Ref skel_type, Skeleton& out) {
         Bone& bo = out.bones[i];
         if (mName)   bo.name   = rd_str(member_slot(b, *mName));
         if (mParent) bo.parent = rd_i32(member_slot(b, *mParent));
-        uint32_t flags = 0;
-        transform_to_composite(member_slot(b, *mLocal), bo.local_transform, flags);
-        bo.lt_flags = flags;
+        read_transform(member_slot(b, *mLocal), bo.local);
+        bo.local_transform = compose(bo.local);
+        bo.lt_flags = bo.local.flags;
         const uint8_t* iw = at(member_slot(b, *mInvW), 64);
         if (iw) for (int k = 0; k < 16; ++k) std::memcpy(&bo.inverse_world[k], iw + k * 4, 4);
         else bo.inverse_world = Mat4{};