Since sort_bounded_heap makes state changes that should be made

regardless of the number of tuples involved, it's incorrect to skip it
when memtupcount = 1; the number of cycles saved is minuscule anyway.
An alternative solution would be to pull the state changes out to the
call site in tuplesort_performsort, but keeping them near the corresponding
changes in make_bounded_heap seems marginally cleaner.  Noticed by
Greg Stark.

diff --git a/src/backend/utils/sort/tuplesort.c b/src/backend/utils/sort/tuplesort.c
index 18e4967d8c660ecfef41724181279e86a460bb29..815dd2158d1507614c334025ef2f83f24f62491e 100644 (file)
--- a/src/backend/utils/sort/tuplesort.c
+++ b/src/backend/utils/sort/tuplesort.c
@@ -91,7 +91,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/utils/sort/tuplesort.c,v 1.77 2007/06/07 19:19:57 tgl Exp $
+ *   $PostgreSQL: pgsql/src/backend/utils/sort/tuplesort.c,v 1.78 2007/09/01 18:47:39 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1068,8 +1068,7 @@ tuplesort_performsort(Tuplesortstate *state)
             * in memory, using a heap to eliminate excess tuples.  Now we have
             * to transform the heap to a properly-sorted array.
             */
-           if (state->memtupcount > 1)
-               sort_bounded_heap(state);
+           sort_bounded_heap(state);
            state->current = 0;
            state->eof_reached = false;
            state->markpos_offset = 0;