fix signed integer overflow in polygon hash calculation

also gives a speedup on some maps lul
This commit is contained in:
Alug committed 2025-11-22 21:47:45 +01:00
1 parent 9a794232d7
commit d46beab34e
1 file changed
+6 -5
+6 -5
View File
@@ -126,8 +126,9 @@ void HWR_ProcessPolygon(FSurfaceInfo *pSurf, FOutVector *pOutVerts, FUINT iNumPt
if (!(PolyFlags & PF_NoTexture) && !horizonSpecial)
{
// use FNV-1a to hash polygons for later sorting.
INT32 hash = 0x811c9dc5;
#define DIGEST(h, x) h ^= (x); h *= 0x01000193
UINT32 hash = 0x811c9dc5;
UINT32 prime = 0x1000193;
#define DIGEST(h, x) h ^= ((UINT32)(x)); h *= prime
if (current_texture)
{
DIGEST(hash, current_texture->downloaded);
@@ -148,7 +149,7 @@ void HWR_ProcessPolygon(FSurfaceInfo *pSurf, FOutVector *pOutVerts, FUINT iNumPt
}
#undef DIGEST
// remove the sign bit to ensure that skybox and horizon line comes first.
polygonArray[polygonArraySize-1].hash = (hash & INT32_MAX);
polygonArray[polygonArraySize-1].hash = (INT32)(hash & INT32_MAX);
}
memcpy(&unsortedVertexArray[unsortedVertexArraySize], pOutVerts, iNumPts * sizeof(FOutVector));
@@ -250,8 +251,6 @@ void HWR_RenderBatches(void)
FSurfaceInfo currentSurfaceInfo;
FSurfaceInfo nextSurfaceInfo;
const boolean useshader = HWR_UseShader();
int i;
if (!currently_batching)
@@ -274,6 +273,8 @@ void HWR_RenderBatches(void)
return;// nothing to draw
}
const boolean useshader = HWR_UseShader();
// init stats vars
ps_hw_numpolys.value.i = polygonArraySize;
ps_hw_numcalls.value.i = ps_hw_numverts.value.i = 0;