nbtree: Demote incomplete split "can't happen" error.

Only a basic logic bug in a _bt_insertonpg() caller could lead to a
violation of this invariant (index corruption won't do it).  A "can't
happen" error seems inappropriate (it is arbitrary at best).

Demote the error to a simple assertion.  This matches similar nearby
sanity checks.

src/backend/access/nbtree/nbtinsert.c

@@ -1129,6 +1129,8 @@ _bt_insertonpg(Relation rel,
 		   IndexRelationGetNumberOfKeyAttributes(rel));
 	Assert(!BTreeTupleIsPosting(itup));
 	Assert(MAXALIGN(IndexTupleSize(itup)) == itemsz);
+	/* Caller must always finish incomplete split for us */
+	Assert(!P_INCOMPLETE_SPLIT(lpageop));
 
 	/*
 	 * Every internal page should have exactly one negative infinity item at
@@ -1138,11 +1140,6 @@ _bt_insertonpg(Relation rel,
 	 */
 	Assert(P_ISLEAF(lpageop) || newitemoff > P_FIRSTDATAKEY(lpageop));
 
-	/* The caller should've finished any incomplete splits already. */
-	if (P_INCOMPLETE_SPLIT(lpageop))
-		elog(ERROR, "cannot insert to incompletely split page %u",
-			 BufferGetBlockNumber(buf));
-
 	/*
 	 * Do we need to split an existing posting list item?
 	 */