Re: GUC for cleanup indexes threshold.

Hi, I tried regression test and found some errors concerning brin and gin,
though I didn't look into this.

Here's a log:

*** /home/ideriha/postgres-master/src/test/regress/expected/brin.out 2017-02-13 11:33:43.270942937 +0900
--- /home/ideriha/postgres-master/src/test/regress/results/brin.out 2017-02-15 14:58:24.725984474 +0900
***************
*** 403,408 ****
SELECT brin_summarize_new_values('brinidx'); -- ok, no change expected
brin_summarize_new_values
---------------------------
! 0
(1 row)

--- 403,408 ----
SELECT brin_summarize_new_values('brinidx'); -- ok, no change expected
brin_summarize_new_values
---------------------------
! 5
(1 row)

======================================================================

*** /home/ideriha/postgres-master/src/test/regress/expected/gin.out 2016-12-20 16:49:09.513050050 +0900
--- /home/ideriha/postgres-master/src/test/regress/results/gin.out 2017-02-15 14:58:25.536984461 +0900
***************
*** 20,26 ****
select gin_clean_pending_list('gin_test_idx'); -- nothing to flush
gin_clean_pending_list
------------------------
! 0
(1 row)

-- Test vacuuming
--- 20,26 ----
select gin_clean_pending_list('gin_test_idx'); -- nothing to flush
gin_clean_pending_list
------------------------
! 8
(1 row)

-- Test vacuuming

======================================================================

Regards,
Ideriha Takeshi