quadtree: fix bad sort due to pointer allocation issue

Author: paulmach

quadtree/maxheap.go

@@ -6,27 +6,18 @@ import "github.com/paulmach/orb"
 // the furthest point from the query point in the list, hence maxHeap.
 // When we find a point closer than the furthest away, we remove
 // furthest and add the new point to the heap.
-type maxHeap []*heapItem
+type maxHeap []heapItem
 
 type heapItem struct {
 	point    orb.Pointer
 	distance float64
 }
 
 func (h *maxHeap) Push(point orb.Pointer, distance float64) {
-	// Common usage is Push followed by a Pop if we have > k points.
-	// We're reusing the k+1 heapItem object to reduce memory allocations.
-	// First we manaully lengthen the slice,
-	// then we see if the last item has been allocated already.
-
 	prevLen := len(*h)
 	*h = (*h)[:prevLen+1]
-	if (*h)[prevLen] == nil {
-		(*h)[prevLen] = &heapItem{point: point, distance: distance}
-	} else {
-		(*h)[prevLen].point = point
-		(*h)[prevLen].distance = distance
-	}
+	(*h)[prevLen].point = point
+	(*h)[prevLen].distance = distance
 
 	i := len(*h) - 1
 	for i > 0 {
@@ -53,21 +44,20 @@ func (h *maxHeap) Push(point orb.Pointer, distance float64) {
 
 // Pop returns the "greatest" item in the list.
 // The returned item should not be saved across push/pop operations.
-func (h *maxHeap) Pop() *heapItem {
-	removed := (*h)[0]
+func (h *maxHeap) Pop() {
 	lastItem := (*h)[len(*h)-1]
 	(*h) = (*h)[:len(*h)-1]
 
 	mh := (*h)
 	if len(mh) == 0 {
-		return removed
+		return
 	}
 
 	// move the last item to the top and reset the heap
-	mh[0] = lastItem
+	mh[0].point = lastItem.point
+	mh[0].distance = lastItem.distance
 
 	i := 0
-	current := mh[i]
 	for {
 		right := (i + 1) << 1
 		left := right - 1
@@ -92,11 +82,12 @@ func (h *maxHeap) Pop() *heapItem {
 		}
 
 		// swap the nodes
-		mh[i] = child
-		mh[childIndex] = current
+		mh[i].point = child.point
+		mh[i].distance = child.distance
+
+		mh[childIndex].point = lastItem.point
+		mh[childIndex].distance = lastItem.distance
 
 		i = childIndex
 	}
-
-	return removed
 }

quadtree/maxheap_test.go

@@ -14,9 +14,11 @@ func TestMaxHeap(t *testing.T) {
 			h.Push(nil, r.Float64())
 		}
 
-		current := h.Pop().distance
+		current := h[0].distance
+		h.Pop()
 		for len(h) > 0 {
-			next := h.Pop().distance
+			next := h[0].distance
+			h.Pop()
 			if next > current {
 				t.Errorf("incorrect")
 			}

quadtree/quadtree.go

@@ -113,6 +113,7 @@ func (q *Quadtree) add(n *node, p orb.Pointer, point orb.Point, left, right, bot
 // Remove will remove the pointer from the quadtree. By default it'll match
 // using the points, but a FilterFunc can be provided for a more specific test
 // if there are elements with the same point value in the tree. For example:
+//
 //	func(pointer orb.Pointer) {
 //		return pointer.(*MyType).ID == lookingFor.ID
 //	}
@@ -273,7 +274,8 @@ func (q *Quadtree) KNearestMatching(buf []orb.Pointer, p orb.Point, k int, f Fil
 	}
 
 	for i := len(v.maxHeap) - 1; i >= 0; i-- {
-		buf[i] = v.maxHeap.Pop().point
+		buf[i] = v.maxHeap[0].point
+		v.maxHeap.Pop()
 	}
 
 	return buf

quadtree/quadtree_test.go

@@ -483,6 +483,27 @@ func TestQuadtreeKNearest_sorted(t *testing.T) {
 	}
 }
 
+func TestQuadtreeKNearest_sorted2(t *testing.T) {
+	q := New(orb.Bound{Max: orb.Point{8, 8}})
+	q.Add(orb.Point{0, 0})
+	q.Add(orb.Point{1, 1})
+	q.Add(orb.Point{2, 2})
+	q.Add(orb.Point{3, 3})
+	q.Add(orb.Point{4, 4})
+	q.Add(orb.Point{5, 5})
+	q.Add(orb.Point{6, 6})
+	q.Add(orb.Point{7, 7})
+
+	nearest := q.KNearest(nil, orb.Point{5.25, 5.25}, 3)
+
+	expected := []orb.Point{{5, 5}, {6, 6}, {4, 4}}
+	for i, p := range expected {
+		if n := nearest[i].Point(); !n.Equal(p) {
+			t.Errorf("incorrect point %d: %v", i, n)
+		}
+	}
+}
+
 func TestQuadtreeKNearest_DistanceLimit(t *testing.T) {
 	type dataPointer struct {
 		orb.Pointer