Clarify why _bt_killitems sorts its items array.

Make it clear why _bt_killitems sorts the scan's so->killedItems[]
array.  Also add an assertion to the _bt_killitems loop (that iterates
through this array) to verify it accesses tuples in leaf page order.

Follow-up to commit bfb335df58.

Author: Peter Geoghegan <pg@bowt.ie>
Suggested-by: Victor Yegorov <vyegorov@gmail.com>
Discussion: https://postgr.es/m/CAGnEboirgArezZDNeFrR8FOGvKF-Xok333s2iVwWi65gZf8MEA@mail.gmail.com

src/backend/access/nbtree/nbtutils.c

@@ -222,11 +222,12 @@ _bt_killitems(IndexScanDesc scan)
 	so->numKilled = 0;
 
 	/*
-	 * so->killedItems[] is in whatever order the scan returned items in.
-	 * Items will appear in descending order during backwards scans.  And
-	 * scrollable cursor scans might have duplicate items.
+	 * We need to iterate through so->killedItems[] in leaf page order; the
+	 * loop below expects this (when marking posting list tuples, at least).
+	 * so->killedItems[] is now in whatever order the scan returned items in.
+	 * Scrollable cursor scans might have even saved the same item/TID twice.
 	 *
-	 * Sort and uniqueify so->killedItems[] to deal with all this.
+	 * Sort and unique-ify so->killedItems[] to deal with all this.
 	 */
 	if (numKilled > 1)
 	{
@@ -271,6 +272,7 @@ _bt_killitems(IndexScanDesc scan)
 	minoff = P_FIRSTDATAKEY(opaque);
 	maxoff = PageGetMaxOffsetNumber(page);
 
+	/* Iterate through so->killedItems[] in leaf page order */
 	for (int i = 0; i < numKilled; i++)
 	{
 		int			itemIndex = so->killedItems[i];
@@ -279,6 +281,9 @@ _bt_killitems(IndexScanDesc scan)
 
 		Assert(itemIndex >= so->currPos.firstItem &&
 			   itemIndex <= so->currPos.lastItem);
+		Assert(i == 0 ||
+			   offnum >= so->currPos.items[so->killedItems[i - 1]].indexOffset);
+
 		if (offnum < minoff)
 			continue;			/* pure paranoia */
 		while (offnum <= maxoff)