To re-iterate my question on Slack from a few days ago, maybe it will have better luck in JIRA:

psergei, is there a reason why "we" keep having these row numbers in MTR test/results?

main.group_min_max 'innodb' w4 [ fail ]

Test ended at 2022-02-09 12:23:42

CURRENT_TEST: main.group_min_max

--- /usr/share/mysql/mysql-test/main/group_min_max.result 2022-02-09 09:34:01.000000000 -0500

+++ /dev/shm/var/4/log/group_min_max.reject 2022-02-09 12:23:41.927795384 -0500

@@ -4059,7 +4059,7 @@

(2, TRUE, "yello"), (2, FALSE, "yello");

EXPLAIN SELECT DISTINCT owner_id FROM t1 WHERE foo = true GROUP BY owner_id HAVING (COUNT(*) = 1);

id select_type table type possible_keys key key_len ref rows Extra

-1 SIMPLE t1 index NULL index_t1_on_owner_id_and_foo 7 NULL 5 Using where; Using index

+1 SIMPLE t1 index NULL index_t1_on_owner_id_and_foo 7 NULL 6 Using where; Using index

SELECT DISTINCT owner_id FROM t1 WHERE foo = true GROUP BY owner_id HAVING (COUNT(*) = 1);

owner_id

1

I understand legacy tests, but why add them in new ones? It's not like anyone ever cares or does anything about these differences

Let me put it this way.
I'm thinking about masking the rows column in every test which fails this way in buildbot. If I do so, will I get any objections from you (or anyone else related)?
And "filtered" too.