because of we divide by /255 not /256, we lose numerical precision, so can't guarantee that box filter results round to same results as naive

tests/resample_test.cpp

@@ -155,7 +155,7 @@ static void performance(int argc, char **argv)
 	out_w = w*2; out_h = h/4; count=20; // 2   // note this is structured pessimily, would be much faster to downsample vertically first
     #elif 0
     out_w = w/4; out_h = h*2; count=50; // 3
-    #elif 0
+    #elif 1
     out_w = w*3; out_h = h*3; count=5; srgb=0; // 4
     #else
     out_w = w*3; out_h = h*3; count=3; // 5   // this is dominated by linear->sRGB conversion
@@ -181,7 +181,7 @@ int main(int argc, char** argv)
 	int out_w, out_h, out_stride;
 
 	//resizer(argc, argv);
-   performance(argc, argv);
+    //performance(argc, argv);
 
 #if 1
 	test_suite(argc, argv);
@@ -648,10 +648,10 @@ void verify_box(void)
 			      + image88[j*2+0][i*2+1]
 				  + image88[j*2+1][i*2+0]
 				  + image88[j*2+1][i*2+1];
-			STBIR_ASSERT(output44[j][i] == ((n+2)>>2));
+			STBIR_ASSERT(output44[j][i] == ((n+2)>>2) || output44[j][i] == ((n+1)>>2)); // can't guarantee exact rounding due to numerical precision
 			t += n;
 		}
-	STBIR_ASSERT(output11[0][0] == ((t+32)>>6));
+	STBIR_ASSERT(output11[0][0] == ((t+32)>>6) || output11[0][0] == ((t+31)>>6)); // can't guarantee exact rounding due to numerical precision
 }
 
 void verify_filter_normalized(stbir_filter filter, int output_size, unsigned int value)
@@ -675,6 +675,8 @@ void test_filters(void)
 {
 	int i,j;
 
+	mtsrand(0);
+
 	for (i=0; i < sizeof(image88); ++i)
 		image88[0][i] = mtrand() & 255;
 	verify_box();