Re: Showing I/O timings spent reading/writing temp buffers in EXPLAIN

Hi,

On Tue, Nov 16, 2021 at 04:37:44PM +0900, Masahiko Sawada wrote:
> 
> I've attached an updated patch. Please review it.

It seems that the regression tests aren't entirely stable, per cfbot:
https://cirrus-ci.com/github/postgresql-cfbot/postgresql/commitfest/36/3298.

The failures look like:

diff -U3 /tmp/cirrus-ci-build/src/test/recovery/../regress/expected/explain.out
/tmp/cirrus-ci-build/src/test/recovery/tmp_check/results/explain.out
--- /tmp/cirrus-ci-build/src/test/recovery/../regress/expected/explain.out    2022-01-19 03:50:37.087931908 +0000
+++ /tmp/cirrus-ci-build/src/test/recovery/tmp_check/results/explain.out    2022-01-19 03:57:41.013616137 +0000
@@ -512,9 +512,10 @@
    I/O Timings: temp read=N.N write=N.N
    ->  Function Scan on generate_series  (cost=N.N..N.N rows=N width=N) (actual time=N.N..N.N rows=N loops=N)
          I/O Timings: temp read=N.N write=N.N
+   I/O Timings: shared/local read=N.N
  Planning Time: N.N ms
  Execution Time: N.N ms
-(6 rows)
+(7 rows)

 select explain_filter('explain (analyze, buffers, format json) select count(*) from generate_series(1, 100000)');
                 explain_filter


I don't see any obvious error in the code, so I'm wondering if it's simply
the result of populating the cache with generate_series() info.