Fix spurious differences when using `ninja diff` (#2294)

* Fix spurious differences when using ninja diff

* Fix some clangd errors/warnings

18 files changed, 98 insertions, 56 deletions

diff --git a/include/JSystem/J2DGraph/J2DAnimation.h b/include/JSystem/J2DGraph/J2DAnimation.h
index cba04c924f..1b4f5c2038 100644
--- a/include/JSystem/J2DGraph/J2DAnimation.h
+++ b/include/JSystem/J2DGraph/J2DAnimation.h
@@ -495,6 +495,7 @@ inline f32 J2DHermiteInterpolation<s16>(register f32 pp1, register s16* pp2, reg
     register s16* p6 = pp6;
     register s16* p7 = pp7;
     // clang-format off
+#ifdef __MWERKS__
     asm {
         psq_l ff2, 0(p2), 0x1, 5
         psq_l ff0, 0(p5), 0x1, 5
@@ -517,6 +518,7 @@ inline f32 J2DHermiteInterpolation<s16>(register f32 pp1, register s16* pp2, reg
         fmadds fout, ff4, ff2, fout
         fsubs fout, fout, ff0
     }
+#endif
     // clang-format on
     return fout;
 }
diff --git a/include/JSystem/J3DGraphAnimator/J3DSkinDeform.h b/include/JSystem/J3DGraphAnimator/J3DSkinDeform.h
index 2cb33d1393..5e03122ce3 100644
--- a/include/JSystem/J3DGraphAnimator/J3DSkinDeform.h
+++ b/include/JSystem/J3DGraphAnimator/J3DSkinDeform.h
@@ -101,6 +101,7 @@ public:
 };  // Size: 0x14
 
 inline void J3DFillZero32B(register void* param_0, register u32 param_1) {
+#ifdef __MWERKS__
     asm {
         srwi param_1, param_1, 5
         mtctr param_1
@@ -109,6 +110,7 @@ inline void J3DFillZero32B(register void* param_0, register u32 param_1) {
     	addi param_0, param_0, 0x20
     	bdnz lbl_8032D948
     }
+#endif
 }
 
 #endif /* J3DSKINDEFORM_H */
diff --git a/include/JSystem/J3DGraphBase/J3DStruct.h b/include/JSystem/J3DGraphBase/J3DStruct.h
index 80f1cdabe1..29fd75b208 100644
--- a/include/JSystem/J3DGraphBase/J3DStruct.h
+++ b/include/JSystem/J3DGraphBase/J3DStruct.h
@@ -56,18 +56,22 @@ struct J3DTextureSRTInfo {
         register const f32* src = &other.mScaleX;
         register f32* dst = &mScaleX;
         register f32 xy;
+#ifdef __MWERKS__
         asm {
             psq_l xy, 0(src), 0, 0
             psq_st xy, 0(dst), 0, 0
         };
+#endif
         // Unclear why there's a 4 byte copy here.
         *(u32*)&mRotation = *(u32*)&other.mRotation;
         src = &other.mTranslationX;
         dst = &mTranslationX;
+#ifdef __MWERKS__
         asm {
             psq_l xy, 0(src), 0, 0
             psq_st xy, 0(dst), 0, 0
         };
+#endif
     }
 };  // Size: 0x14
 
diff --git a/include/JSystem/J3DGraphBase/J3DTevs.h b/include/JSystem/J3DGraphBase/J3DTevs.h
index 136df1f7a5..f25c0a1a11 100644
--- a/include/JSystem/J3DGraphBase/J3DTevs.h
+++ b/include/JSystem/J3DGraphBase/J3DTevs.h
@@ -240,14 +240,14 @@ struct J3DTevOrder : public J3DTevOrderInfo {
 };
 
 extern u8 j3dTevSwapTableTable[1024];
-extern u8 const data_804563F8;
+extern u8 const j3dDefaultTevSwapTableID;
 
 /**
  * @ingroup jsystem-j3d
  * 
  */
 struct J3DTevSwapModeTable {
-    /* 8000E134 */ J3DTevSwapModeTable() { mIdx = data_804563F8; }
+    /* 8000E134 */ J3DTevSwapModeTable() { mIdx = j3dDefaultTevSwapTableID; }
     J3DTevSwapModeTable(J3DTevSwapModeTableInfo const& info) {
         mIdx = calcTevSwapTableID(info.field_0x0, info.field_0x1, info.field_0x2, info.field_0x3);
     }
diff --git a/include/JSystem/J3DGraphBase/J3DTransform.h b/include/JSystem/J3DGraphBase/J3DTransform.h
index 9225c44605..28b79d8a9a 100644
--- a/include/JSystem/J3DGraphBase/J3DTransform.h
+++ b/include/JSystem/J3DGraphBase/J3DTransform.h
@@ -76,6 +76,7 @@ inline void J3DPSMtx33Copy(register Mtx3P src, register Mtx3P dst) {
     register f32 fr2;
     register f32 fr1;
     register f32 fr0;
+#ifdef __MWERKS__
     asm {
         psq_l fr4, 0(src), 0, 0
         psq_l fr3, 8(src), 0, 0
@@ -88,6 +89,7 @@ inline void J3DPSMtx33Copy(register Mtx3P src, register Mtx3P dst) {
         psq_st fr1, 0x18(dst), 0, 0
         stfs fr0, 0x20(dst)
     }
+#endif
 }
 
 inline void J3DPSMtx33CopyFrom34(register MtxP src, register Mtx3P dst) {
@@ -97,6 +99,7 @@ inline void J3DPSMtx33CopyFrom34(register MtxP src, register Mtx3P dst) {
     register f32 z2;
     register f32 x_y3;
     register f32 z3;
+#ifdef __MWERKS__
     asm {
         psq_l x_y1, 0(src), 0, 0
         lfs z1, 8(src)
@@ -111,6 +114,7 @@ inline void J3DPSMtx33CopyFrom34(register MtxP src, register Mtx3P dst) {
         psq_st x_y3, 24(dst), 0, 0
         stfs z3, 0x20(dst)
     }
+#endif
 }
 
 // regalloc issues
@@ -131,6 +135,7 @@ inline void J3DPSMulMtxVec(register MtxP mtx, register Vec* vec, register Vec* d
     register f32 fra2;
     register f32 fr01;
     register f32 fr00;
+#ifdef __MWERKS__
     asm {
         psq_l fr00, 0(vec), 0, 0
     	psq_l fr2, 0(mtx), 0, 0
@@ -153,6 +158,7 @@ inline void J3DPSMulMtxVec(register MtxP mtx, register Vec* vec, register Vec* d
     	ps_sum0 fra6, fra5, fra6, fra5
     	psq_st fra6, 8(dst), 1, 0
     }
+#endif
 }
 
 // regalloc issues
@@ -173,6 +179,7 @@ inline void J3DPSMulMtxVec(register MtxP mtx, register SVec* vec, register SVec*
     register f32 fra2;
     register f32 fr01;
     register f32 fr00;
+#ifdef __MWERKS__
     asm {
         psq_l fr00, 0(vec), 0, 7
     	psq_l fr2, 0(mtx), 0, 0
@@ -195,6 +202,7 @@ inline void J3DPSMulMtxVec(register MtxP mtx, register SVec* vec, register SVec*
     	ps_sum0 fra6, fra5, fra6, fra5
     	psq_st fra6, 4(dst), 1, 7
     }
+#endif
 }
 
 // regalloc issues
@@ -213,6 +221,7 @@ inline void J3DPSMulMtxVec(register Mtx3P mtx, register Vec* vec, register Vec*
     register f32 fr2;
     register f32 fr01;
     register f32 fr00;
+#ifdef __MWERKS__
     asm {
         lis punit, PSMulUnit01@ha
         psq_l fr00, 0(vec), 0, 0
@@ -239,6 +248,7 @@ inline void J3DPSMulMtxVec(register Mtx3P mtx, register Vec* vec, register Vec*
         ps_sum0 fr6, fr5, fr6, fr5
         psq_st fr6, 8(dst), 1, 0
     }
+#endif
 }
 
 // regalloc issues
@@ -252,6 +262,7 @@ inline void J3DPSMulMtxVec(register Mtx3P mtx, register SVec* vec, register SVec
     register f32 fr2;
     register f32 fr01;
     register f32 fr00;
+#ifdef __MWERKS__
     asm {
         lis punit, PSMulUnit01@ha
         psq_l fr00, 0(vec), 0, 7
@@ -278,6 +289,7 @@ inline void J3DPSMulMtxVec(register Mtx3P mtx, register SVec* vec, register SVec
         ps_sum0 fr6, fr5, fr6, fr5
         psq_st fr6, 4(dst), 1, 7
     }
+#endif
 }
 
 #endif /* J3DTRANSFORM_H */
diff --git a/include/JSystem/JAudio2/JASDSPInterface.h b/include/JSystem/JAudio2/JASDSPInterface.h
index 996053975b..370fe738e4 100644
--- a/include/JSystem/JAudio2/JASDSPInterface.h
+++ b/include/JSystem/JAudio2/JASDSPInterface.h
@@ -114,7 +114,7 @@ namespace JASDsp {
 
     extern u8 const DSPADPCM_FILTER[64];
     extern u32 const DSPRES_FILTER[320];
-    extern u16 SEND_TABLE[12 + 12 /* padding */];
+    extern u16 SEND_TABLE[];
     extern TChannel* CH_BUF;
     extern FxBuf* FX_BUF;
     extern f32 sDSPVolume;
diff --git a/include/JSystem/JAudio2/JASHeapCtrl.h b/include/JSystem/JAudio2/JASHeapCtrl.h
index 9c174511ae..5eae318da5 100644
--- a/include/JSystem/JAudio2/JASHeapCtrl.h
+++ b/include/JSystem/JAudio2/JASHeapCtrl.h
@@ -99,6 +99,8 @@ public:
     void free(void* ptr, u32 n) { JASGenericMemPool::free(ptr, n); }
 };
 
+namespace JASKernel { JKRHeap* getSystemHeap(); };
+
 /**
  * @ingroup jsystem-jaudio
  * 
@@ -183,7 +185,7 @@ public:
     }
 
     void* alloc(u32 size) {
-        T::Lock lock(&mMutex);
+        typename T::Lock lock(&mMutex);
         if (field_0x18->getFreeSize() < size) {
             if (ChunkSize < size) {
                 return NULL;
@@ -196,7 +198,7 @@ public:
     }
 
     void free(void* ptr) {
-        T::Lock lock(&mMutex);
+        typename T::Lock lock(&mMutex);
         MemoryChunk* chunk = field_0x18;
         MemoryChunk* prevChunk = NULL;
         while (chunk != NULL) {
@@ -269,17 +271,17 @@ template <typename T>
 class JASMemPool_MultiThreaded : public JASGenericMemPool {
 public:
     void newMemPool(int param_0) {
-        JASThreadingModel::InterruptsDisable<JASMemPool_MultiThreaded<T> >::Lock lock(*this);
+        typename JASThreadingModel::InterruptsDisable<JASMemPool_MultiThreaded<T> >::Lock lock(*this);
         JASGenericMemPool::newMemPool(sizeof(T), param_0);
     }
 
     void* alloc(size_t count) {
-        JASThreadingModel::InterruptsDisable<JASMemPool_MultiThreaded<T> >::Lock lock(*this);
+        typename JASThreadingModel::InterruptsDisable<JASMemPool_MultiThreaded<T> >::Lock lock(*this);
         return JASGenericMemPool::alloc(count);
     }
 
     void free(void* ptr, u32 param_1) {
-        JASThreadingModel::InterruptsDisable<JASMemPool_MultiThreaded<T> >::Lock lock(*this);
+        typename JASThreadingModel::InterruptsDisable<JASMemPool_MultiThreaded<T> >::Lock lock(*this);
         JASGenericMemPool::free(ptr, param_1);
     }
 };
diff --git a/include/JSystem/JGeometry.h b/include/JSystem/JGeometry.h
index 141b672366..281f7d5b40 100644
--- a/include/JSystem/JGeometry.h
+++ b/include/JSystem/JGeometry.h
@@ -117,12 +117,14 @@ inline void setTVec3f(const f32* vec_a, f32* vec_b) {
     register f32 a_x;
     register f32 b_x;
 
+#ifdef __MWERKS__
     asm {
         psq_l a_x, 0(v_a), 0, 0
         lfs b_x, 8(v_a)
         psq_st a_x, 0(v_b), 0, 0
         stfs b_x, 8(v_b)
     };
+#endif
 }
 
 // Until we figure out TVec3 ctors
@@ -188,6 +190,7 @@ struct TVec3<f32> : public Vec {
         register f32 za;
         register f32 zb;
         register f32 z;
+#ifdef __MWERKS__
         asm {
             psq_l  a_x_y, 0(srca), 0, 0
             psq_l  b_x_y, 0(srcb), 0, 0
@@ -198,6 +201,7 @@ struct TVec3<f32> : public Vec {
             fmuls  z, za, zb
             stfs   z,   8(dst)
         };
+#endif
     }
 
     inline void mul(const TVec3<f32>& a) {
@@ -273,6 +277,7 @@ struct TVec3<f32> : public Vec {
         register f32 x_y;
         register f32* dst = &x;
         register f32 zres;
+#ifdef __MWERKS__
         asm {
             psq_l    x_y, 0(dst),  0, 0
             psq_l    z,   8(dst),  1, 0
@@ -281,6 +286,7 @@ struct TVec3<f32> : public Vec {
             ps_muls0 zres,       z, sc
             psq_st   zres,  8(dst),  1, 0
         };
+#endif
     }
 
     void scale(register f32 sc, const TVec3<f32>& other) {
@@ -289,6 +295,7 @@ struct TVec3<f32> : public Vec {
         register f32 x_y;
         register f32* dst = &x;
         register f32 zres;
+#ifdef __MWERKS__
         asm {
             psq_l    x_y, 0(src),  0, 0
             psq_l    z,   8(src),  1, 0
@@ -297,6 +304,7 @@ struct TVec3<f32> : public Vec {
             ps_muls0 zres,       z, sc
             psq_st   zres,  8(dst),  1, 0
         };
+#endif
     }
 
     void scaleAdd(register f32 sc, const TVec3<f32>& a, const TVec3<f32>& b) {
@@ -308,6 +316,7 @@ struct TVec3<f32> : public Vec {
         const register f32* src = &x;
         register f32 x_y;
         register f32 z;
+#ifdef __MWERKS__
         asm {
             psq_l  x_y, 0(src), 0, 0
             ps_neg x_y, x_y
@@ -316,6 +325,7 @@ struct TVec3<f32> : public Vec {
             fneg   z,   z
             stfs   z,   8(rdst)
         };
+#endif
     }
 
     void negate() {
@@ -371,6 +381,7 @@ struct TVec3<f32> : public Vec {
         register f32 otheryz;
         register f32 otherxy;
         register f32 thisxy;
+#ifdef __MWERKS__
         asm {
             psq_l thisyz, 4(pThis), 0, 0
             psq_l otheryz, 4(pOther), 0, 0
@@ -380,6 +391,7 @@ struct TVec3<f32> : public Vec {
             ps_madd otheryz, thisxy, otherxy, thisyz
             ps_sum0 res, otheryz, thisyz, thisyz
         };
+#endif
         return res;
     }
 
@@ -462,7 +474,7 @@ struct TVec2 {
 template <class T>
 struct TBox {
     TBox() : i(), f() {}
-    TBox(const TBox& other) : i(other.f), f(other.y) {}
+    TBox(const TBox& other) : i(other.i), f(other.f) {}
 
     T i, f;
 };
@@ -510,7 +522,7 @@ struct TBox2 : TBox<TVec2<T> > {
 
     void set(const TBox2& other) { set(other.i, other.f); }
     void set(const TVec2<f32>& i, const TVec2<f32>& f) { this->i.set(i), this->f.set(f); }
-    void set(f32 x0, f32 y0, f32 x1, f32 y1) { i.set(x0, y0); f.set(x1, y1); }
+    void set(f32 x0, f32 y0, f32 x1, f32 y1) { this->i.set(x0, y0); this->f.set(x1, y1); }
 };
 
 // clang-format on
diff --git a/include/JSystem/JMath/JMath.h b/include/JSystem/JMath/JMath.h
index e37460e115..0968feed00 100644
--- a/include/JSystem/JMath/JMath.h
+++ b/include/JSystem/JMath/JMath.h
@@ -2,6 +2,7 @@
 #define JMATH_H
 
 #include "dolphin/mtx.h"
+#include "math.h"
 
 void JMAMTXApplyScale(const Mtx, Mtx, f32, f32, f32);
 void JMAEulerToQuat(s16 param_0, s16 param_1, s16 param_2, Quaternion* param_3);
@@ -21,9 +22,11 @@ inline f32 JMAFastReciprocal(f32 value) {
 inline float __frsqrtes(register double f) {
     register float out;
     // clang-format off
+#ifdef __MWERKS__
     asm {
         frsqrte out, f
     }
+#endif
     // clang-format on
     return out;
 }
@@ -31,9 +34,11 @@ inline float __frsqrtes(register double f) {
 inline f32 JMAFastSqrt(register f32 input) {
     if (input > 0.0f) {
         register f32 out;
+#ifdef __MWERKS__
         asm {
             frsqrte out, input
         }
+#endif
         return out * input;
     } else {
         return input;
@@ -51,6 +56,7 @@ inline f32 JMAHermiteInterpolation(register f32 p1, register f32 p2, register f3
     register f32 ff27;
     register f32 ff26;
     // clang-format off
+#ifdef __MWERKS__
     asm {
         fsubs   ff31, p1, p2
         fsubs   ff30, p5, p2
@@ -67,6 +73,7 @@ inline f32 JMAHermiteInterpolation(register f32 p1, register f32 p2, register f3
         fnmsubs ff25,ff31,ff25,ff26
 
     }
+#endif
     // clang-format on
     return ff25;
 }
@@ -80,18 +87,21 @@ inline f32 fastReciprocal(f32 value) {
 inline void gekko_ps_copy3(register void* dst, register const void* src) {
     register f32 src0;
     register f32 src1;
+#ifdef __MWERKS__
     asm {
         psq_l src0, 0(src), 0, 0
         lfs src1, 8(src)
         psq_st src0, 0(dst), 0, 0
         stfs src1, 8(dst)
     };
+#endif
 }
 
 inline void gekko_ps_copy6(register void* dst, register const void* src) {
     register f32 src0;
     register f32 src1;
     register f32 src2;
+#ifdef __MWERKS__
     asm {
         psq_l src0, 0(src), 0, 0
         psq_l src1, 8(src), 0, 0
@@ -100,6 +110,7 @@ inline void gekko_ps_copy6(register void* dst, register const void* src) {
         psq_st src1, 8(dst), 0, 0
         psq_st src2, 16(dst), 0, 0
     };
+#endif
 }
 
 inline void gekko_ps_copy12(register void* dst, register const void* src) {
@@ -109,6 +120,7 @@ inline void gekko_ps_copy12(register void* dst, register const void* src) {
     register f32 src3;
     register f32 src4;
     register f32 src5;
+#ifdef __MWERKS__
     asm {
         psq_l src0, 0(src), 0, 0
         psq_l src1, 8(src), 0, 0
@@ -123,6 +135,7 @@ inline void gekko_ps_copy12(register void* dst, register const void* src) {
         psq_st src4, 32(dst), 0, 0
         psq_st src5, 40(dst), 0, 0
     };
+#endif
 }
 
 inline void gekko_ps_copy16(register void* dst, register const void* src) {
@@ -134,6 +147,7 @@ inline void gekko_ps_copy16(register void* dst, register const void* src) {
     register f32 src5;
     register f32 src6;
     register f32 src7;
+#ifdef __MWERKS__
     asm {
         psq_l src0, 0(src), 0, 0
         psq_l src1, 8(src), 0, 0
@@ -152,6 +166,7 @@ inline void gekko_ps_copy16(register void* dst, register const void* src) {
         psq_st src6, 48(dst), 0, 0
         psq_st src7, 56(dst), 0, 0
     };
+#endif
 }
 
 };  // namespace JMath
diff --git a/include/JSystem/JMessage/resource.h b/include/JSystem/JMessage/resource.h
index a322c5b952..4a4b990dc3 100644
--- a/include/JSystem/JMessage/resource.h
+++ b/include/JSystem/JMessage/resource.h
@@ -114,7 +114,7 @@ struct TResourceContainer {
     static JMessage::locale::parseCharacter_function sapfnParseCharacter_[5];
 
     /* 0x00 */ u8 mEncodingType;
-    /* 0x04 */ int (*pfnParseCharacter_)(const char** string) const;
+    /* 0x04 */ int (*pfnParseCharacter_)(const char** string);
     /* 0x08 */ TCResource resContainer_;
     /* 0x18 */ TResource_color resColor_;
 };
diff --git a/include/SSystem/SComponent/c_lib.h b/include/SSystem/SComponent/c_lib.h
index c6ee07af1f..99693a0305 100644
--- a/include/SSystem/SComponent/c_lib.h
+++ b/include/SSystem/SComponent/c_lib.h
@@ -3,6 +3,7 @@
 
 #include "SSystem/SComponent/c_xyz.h"
 #include "dolphin/mtx.h"
+#include "SSystem/SComponent/c_math.h"
 
 inline bool cLib_IsZero(f32 value) {
     return fabsf(value) < 8e-11f;
diff --git a/include/d/d_bright_check.h b/include/d/d_bright_check.h
index bab911da21..53c0962919 100644
--- a/include/d/d_bright_check.h
+++ b/include/d/d_bright_check.h
@@ -25,12 +25,6 @@ public:
     /* 80192F10 */ dBrightCheck_c(JKRArchive*);
     /* 80192F98 */ virtual ~dBrightCheck_c();
 
-    // TODO: probably fake, probably going to break some other function
-    // need to figure out this weird vtable padding issue
-    virtual void dummy() = 0;
-    virtual void dummy2() = 0;
-    virtual void dummy3() = 0;
-
     /* 80193030 */ void screenSet();
     /* 801934D0 */ void _move();
     /* 80193508 */ void modeWait();
diff --git a/include/d/d_file_sel_info.h b/include/d/d_file_sel_info.h
index ded203c308..0d75e7f9d6 100644
--- a/include/d/d_file_sel_info.h
+++ b/include/d/d_file_sel_info.h
@@ -32,12 +32,6 @@ public:
 
     /* 801924A0 */ virtual ~dFile_info_c();
 
-    // fake? needed to get vtable size correct
-    #ifdef DFILE_INFO_C_DUMMY_VIRTUAL
-    virtual void dummy() = 0;
-    virtual void dummy2() = 0;
-    #endif
-
     void setBasePane(J2DPane* pane) { mFileInfo.mBasePane = pane; }
     CPaneMgrAlpha* getDatBase() { return mDatBase; }
     CPaneMgrAlpha* getNoDatBase() { return mNoDatBase; }
diff --git a/include/d/d_file_sel_warning.h b/include/d/d_file_sel_warning.h
index 8164ed4afd..640f26e99c 100644
--- a/include/d/d_file_sel_warning.h
+++ b/include/d/d_file_sel_warning.h
@@ -38,12 +38,6 @@ public:
 
     /* 80191C18 */ virtual ~dFile_warning_c();
 
-    // fake? needed to get vtable size correct
-    #ifdef DFILE_WARNING_C_DUMMY_VIRTUAL
-    virtual void dummy() = 0;
-    virtual void dummy2() = 0;
-    #endif
-
     u8 getStatus() { return mStatus; }
     void draw() { _draw(); }
 
diff --git a/include/d/d_menu_calibration.h b/include/d/d_menu_calibration.h
index 4498fa3195..f8abc508b3 100644
--- a/include/d/d_menu_calibration.h
+++ b/include/d/d_menu_calibration.h
@@ -43,11 +43,6 @@ public:
     /* 801AFBF4 */ virtual void draw();
     /* 801AED64 */ virtual ~dMenu_Calibration_c();
 
-    // this is just a hack to get .data section to match
-    // not sure how to fix it without this atm
-    virtual void unk0() = 0;
-    virtual void unk1() = 0;
-
     /* 0x004 */ JKRExpHeap* mpHeap;
     /* 0x008 */ JKRArchive* mpArchive;
     /* 0x00C */ J2DScreen* mpCalibrationMainScrn;
diff --git a/include/dolphin/mtx/vec.h b/include/dolphin/mtx/vec.h
index 91216e7464..d3f7a8e365 100644
--- a/include/dolphin/mtx/vec.h
+++ b/include/dolphin/mtx/vec.h
@@ -39,6 +39,7 @@ inline void C_VECAdd(register const Vec* a, register const Vec* b, register Vec*
     register f32 az;
     register f32 sumz;
     register f32 bz;
+#ifdef __MWERKS__
     asm {
         psq_l axy, 0(a), 0, 0
         psq_l bxy, 0(b), 0, 0
@@ -49,6 +50,7 @@ inline void C_VECAdd(register const Vec* a, register const Vec* b, register Vec*
         ps_add sumz, az, bz
         psq_st sumz, 8(ab), 1, 0
     }
+#endif
 }
 
 inline void C_VECSubtract(register const Vec* a, register const Vec* b, register Vec* ab) {
@@ -57,6 +59,7 @@ inline void C_VECSubtract(register const Vec* a, register const Vec* b, register
     register f32 az;
     register f32 subz;
     register f32 bz;
+#ifdef __MWERKS__
     asm {
         psq_l axy, 0(a), 0, 0
         psq_l bxy, 0(b), 0, 0
@@ -67,6 +70,7 @@ inline void C_VECSubtract(register const Vec* a, register const Vec* b, register
         ps_sub subz, az, bz
         psq_st subz, 8(ab), 1, 0
     }
+#endif
 }
 
 inline f32 C_VECSquareMag(const Vec* v) {
@@ -74,6 +78,7 @@ inline f32 C_VECSquareMag(const Vec* v) {
     register f32 z;
     register f32 res;
     register const f32* src = &v->x;
+#ifdef __MWERKS__
     asm {
         psq_l   x_y, 0(src), 0, 0
         ps_mul  x_y, x_y, x_y
@@ -81,7 +86,7 @@ inline f32 C_VECSquareMag(const Vec* v) {
         ps_madd res, z, z, x_y
         ps_sum0 res, res, x_y, x_y
     }
-    ;
+#endif
     return res;
 }
 
diff --git a/include/dolphin/os.h b/include/dolphin/os.h
index 1537226f90..0fbf83fce4 100644
--- a/include/dolphin/os.h
+++ b/include/dolphin/os.h
@@ -5,27 +5,27 @@
 #include "macros.h"
 #include "dolphin/dvd.h"
 
-#include "dolphin/os/OSAlarm.h"
-#include "dolphin/os/OSAlloc.h"
-#include "dolphin/os/OSArena.h"
-#include "dolphin/os/OSAudioSystem.h"
-#include "dolphin/os/OSCache.h"
-#include "dolphin/os/OSContext.h"
-#include "dolphin/os/OSError.h"
-#include "dolphin/os/OSExec.h"
-#include "dolphin/os/OSFont.h"
-#include "dolphin/os/OSInterrupt.h"
-#include "dolphin/os/OSLink.h"
-#include "dolphin/os/OSMemory.h"
-#include "dolphin/os/OSMessage.h"
-#include "dolphin/os/OSMutex.h"
-#include "dolphin/os/OSReboot.h"
-#include "dolphin/os/OSReset.h"
-#include "dolphin/os/OSResetSW.h"
-#include "dolphin/os/OSRtc.h"
-#include "dolphin/os/OSSync.h"
-#include "dolphin/os/OSThread.h"
-#include "dolphin/os/OSTime.h"
+#include "dolphin/os/OSAlarm.h" // IWYU pragma: export
+#include "dolphin/os/OSAlloc.h" // IWYU pragma: export
+#include "dolphin/os/OSArena.h" // IWYU pragma: export
+#include "dolphin/os/OSAudioSystem.h" // IWYU pragma: export
+#include "dolphin/os/OSCache.h" // IWYU pragma: export
+#include "dolphin/os/OSContext.h" // IWYU pragma: export
+#include "dolphin/os/OSError.h" // IWYU pragma: export
+#include "dolphin/os/OSExec.h" // IWYU pragma: export
+#include "dolphin/os/OSFont.h" // IWYU pragma: export
+#include "dolphin/os/OSInterrupt.h" // IWYU pragma: export
+#include "dolphin/os/OSLink.h" // IWYU pragma: export
+#include "dolphin/os/OSMemory.h" // IWYU pragma: export
+#include "dolphin/os/OSMessage.h" // IWYU pragma: export
+#include "dolphin/os/OSMutex.h" // IWYU pragma: export
+#include "dolphin/os/OSReboot.h" // IWYU pragma: export
+#include "dolphin/os/OSReset.h" // IWYU pragma: export
+#include "dolphin/os/OSResetSW.h" // IWYU pragma: export
+#include "dolphin/os/OSRtc.h" // IWYU pragma: export
+#include "dolphin/os/OSSync.h" // IWYU pragma: export
+#include "dolphin/os/OSThread.h" // IWYU pragma: export
+#include "dolphin/os/OSTime.h" // IWYU pragma: export
 
 void OSReportInit(void);
 void OSSwitchFiberEx(u32, u32, u32, u32, u32, u32);
@@ -145,10 +145,12 @@ inline s16 __OSf32tos16(register f32 inF) {
     u32 tmp;
     register u32* tmpPtr = &tmp;
     // clang-format off
+#ifdef __MWERKS__
     asm {
         psq_st inF, 0(tmpPtr), 0x1, 5
         lha out, 0(tmpPtr)
     }
+#endif
     // clang-format on
 
     return out;
@@ -163,10 +165,12 @@ inline u8 __OSf32tou8(register f32 inF) {
     u32 tmp;
     register u32* tmpPtr = &tmp;
     // clang-format off
+#ifdef __MWERKS__
     asm {
         psq_st inF, 0(tmpPtr), 0x1, 2
         lbz out, 0(tmpPtr)
     }
+#endif
     // clang-format on
 
     return out;
@@ -178,6 +182,7 @@ inline void OSf32tou8(f32* f, u8* out) {
 
 static inline void OSInitFastCast(void) {
     // clang-format off
+#ifdef __MWERKS__
     asm {
         li r3, 4
         oris r3, r3, 4
@@ -192,6 +197,7 @@ static inline void OSInitFastCast(void) {
         oris r3, r3, 7
         mtspr 0x395, r3
     }
+#endif
     // clang-format on
 }
 
diff --git a/include/global.h b/include/global.h
index d626b182ab..5cb534ede8 100644
--- a/include/global.h
+++ b/include/global.h
@@ -48,6 +48,10 @@ inline BOOL checkEqual(s32 a, s32 b) {
     return (u32)__cntlzw(a - b) >> 5;
 }
 
+#ifndef __MWERKS__
+void* __memcpy(void*, const void*, int);
+#endif
+
 #define FAST_DIV(x, n) (x >> (n / 2))
 
 #define SQUARE(x) ((x) * (x))