Decompile WRP func_801907EC (!) (#1322)

One of the last two functions!

Since this is duplicated in every overlay (but never used in any of
them), it will make sense to de-duplicate this. I will use my automatic
de-duplication script to put it into the appropriate .h file. That can
either be in this PR (after getting the seal of approval), or can be a
separate PR (after merging this initial WRP version). I'm good with any
option.

Very cool though, glad I figured this one out finally! Hopefully the
last one will come along soon...

include/types.h

@@ -53,4 +53,9 @@ typedef struct {
     /* 0x2 */ s16 y;
 } Point16; // size = 0x4
 
+typedef struct {
+    u8 u;
+    u8 v;
+} uvPair;
+
 #endif

src/pc/psxsdk/libgte.c

@@ -78,3 +78,15 @@ long RotAverage3(SVECTOR* v0, SVECTOR* v1, SVECTOR* v2, long* sxy0, long* sxy1,
     NOT_IMPLEMENTED;
     return 0;
 }
+
+long RotTransPers4(
+    SVECTOR* v0, SVECTOR* v1, SVECTOR* v2,
+    SVECTOR* v3, // Pointers to vectors (input)
+    long* v10, long* v11, long* v12,
+    long* v13, // Pointers to screen coordinates
+    long* p,   // Pointer to interpolated value for depth cueing
+    long* flag // Pointer to flag)
+) {
+    NOT_IMPLEMENTED;
+    return 0;
+}

src/st/wrp/e_misc.c

@@ -410,7 +410,111 @@ extern u8 g_bigRedFireballAnim[];
 u16 D_8018105C[] = {0, 1, 3, 4, 1, 2, 4, 5, 3, 4, 6, 7, 4, 5, 7, 8};
 #include "../entity_big_red_fireball.h"
 
-INCLUDE_ASM("st/wrp/nonmatchings/e_misc", func_801907EC);
+// arg7 holds, in order:
+// a Primitive, an array of 9 SVECTORs, an array of 10 uvPairs, and a pointer to
+// another of arg7.
+// arg7 is not actually u8* but windows doesn't like void* I guess?
+Primitive* func_801907EC(
+    SVECTOR* arg0, SVECTOR* arg1, SVECTOR* arg2, SVECTOR* arg3, Primitive* arg4,
+    s32 arg5, Primitive* arg6, u8* arg7) {
+    s32 unused1;
+    s32 unused2;
+    s32 i;
+    Primitive* prim1;
+    Primitive* prim2;
+    s32 vecIdx1;
+    s32 vecIdx2;
+    s32 vecIdx3;
+    s32 vecIdx4;
+    s16* vecIndices;
+    s32 rotTransResult;
+    SVECTOR* vectors;
+    uvPair* uv_values;
+    Primitive* nextArg6;
+
+    if (arg6 == NULL) {
+        return NULL;
+    }
+    prim2 = arg7;
+    arg7 += 0x34;
+    vectors = arg7;
+    arg7 += 0x48;
+
+    vectors[0] = *arg0;
+    vectors[2] = *arg1;
+    vectors[6] = *arg2;
+    vectors[8] = *arg3;
+
+    uv_values = arg7;
+    arg7 += 0x14; // This is effectively arg7 = arg7->next, since this covers
+                  // the full size of arg7.
+
+    prim1 = prim2;
+    vectors[1].vx = (vectors[0].vx + vectors[2].vx + 1) >> 1;
+    vectors[1].vy = (vectors[0].vy + vectors[2].vy + 1) >> 1;
+    vectors[1].vz = (vectors[0].vz + vectors[2].vz + 1) >> 1;
+    vectors[7].vx = (vectors[6].vx + vectors[8].vx + 1) >> 1;
+    vectors[7].vy = (vectors[6].vy + vectors[8].vy + 1) >> 1;
+    vectors[7].vz = (vectors[6].vz + vectors[8].vz + 1) >> 1;
+    vectors[3].vx = (vectors[0].vx + vectors[6].vx + 1) >> 1;
+    vectors[3].vy = (vectors[0].vy + vectors[6].vy + 1) >> 1;
+    vectors[3].vz = (vectors[0].vz + vectors[6].vz + 1) >> 1;
+    vectors[5].vx = (vectors[2].vx + vectors[8].vx + 1) >> 1;
+    vectors[5].vy = (vectors[2].vy + vectors[8].vy + 1) >> 1;
+    vectors[5].vz = (vectors[2].vz + vectors[8].vz + 1) >> 1;
+    vectors[4].vx = (vectors[3].vx + vectors[5].vx + 1) >> 1;
+    vectors[4].vy = (vectors[3].vy + vectors[5].vy + 1) >> 1;
+    vectors[4].vz = (vectors[3].vz + vectors[5].vz + 1) >> 1;
+
+    uv_values[0] = *(uvPair*)&arg4->u0;
+    uv_values[2] = *(uvPair*)&arg4->u1;
+    uv_values[6] = *(uvPair*)&arg4->u2;
+    uv_values[8] = *(uvPair*)&arg4->u3;
+    uv_values[1].u = (uv_values[0].u + uv_values[2].u + 1) >> 1;
+    uv_values[1].v = (uv_values[0].v + uv_values[2].v + 1) >> 1;
+    uv_values[7].u = (uv_values[6].u + uv_values[8].u + 1) >> 1;
+    uv_values[7].v = (uv_values[6].v + uv_values[8].v + 1) >> 1;
+    uv_values[3].u = (uv_values[0].u + uv_values[6].u + 1) >> 1;
+    uv_values[3].v = (uv_values[0].v + uv_values[6].v + 1) >> 1;
+    uv_values[5].u = (uv_values[2].u + uv_values[8].u + 1) >> 1;
+    uv_values[5].v = (uv_values[2].v + uv_values[8].v + 1) >> 1;
+    uv_values[4].u = (uv_values[3].u + uv_values[5].u + 1) >> 1;
+    uv_values[4].v = (uv_values[3].v + uv_values[5].v + 1) >> 1;
+
+    *prim1 = *arg4;
+    vecIndices = &D_8018105C;
+    for (i = 0; i < 4; i++) {
+        vecIdx1 = *vecIndices++;
+        vecIdx2 = *vecIndices++;
+        vecIdx3 = *vecIndices++;
+        vecIdx4 = *vecIndices++;
+        rotTransResult = RotTransPers4(
+            &vectors[vecIdx1], &vectors[vecIdx2], &vectors[vecIdx3],
+            &vectors[vecIdx4], &prim2->x0, &prim2->x1, &prim2->x2, &prim2->x3,
+            &unused1, &unused2);
+        *(uvPair*)&prim2->u0 = uv_values[vecIdx1];
+        *(uvPair*)&prim2->u1 = uv_values[vecIdx2];
+        *(uvPair*)&prim2->u2 = uv_values[vecIdx3];
+        *(uvPair*)&prim2->u3 = uv_values[vecIdx4];
+        if (arg5 == 1) {
+            prim1 = arg6;
+            if (rotTransResult > 0) {
+                nextArg6 = arg6->next;
+                *prim1 = *prim2;
+                arg6->next = nextArg6;
+                arg6 = nextArg6;
+                if (arg6 == NULL) {
+                    return NULL;
+                }
+            }
+        } else {
+            arg6 = func_801907EC(
+                &vectors[vecIdx1], &vectors[vecIdx2], &vectors[vecIdx3],
+                &vectors[vecIdx4], prim2, arg5 - 1, arg6, arg7);
+        }
+    }
+    return arg6;
+}
 
 u16 D_8018107C[] = {0, 1, 3, 4, 1, 2, 4, 5, 3, 4, 6, 7, 4, 5, 7, 8, 0, 0};
 INCLUDE_ASM("st/wrp/nonmatchings/e_misc", func_80190DCC);