GROUP BY-subquery crashes MonetDb

[jniewerth]

Describe the bug
In a query that has a subquery with a GROUP BY-clause in the FROM clause MonetDB server crashes when having another correlated subquery in the SELECT clause.

To Reproduce

root@dd2a1e8e05d4:~# mclient testdb
Welcome to mclient, the MonetDB/SQL interactive terminal (unreleased)
Database: MonetDB v11.39.18 (hg id: d0fc592188), 'mapi:monetdb://dd2a1e8e05d4:50000/testdb'
FOLLOW US on https://twitter.com/MonetDB or https://github.com/MonetDB/MonetDB
Type \q to quit, \? for a list of available commands
auto commit mode: on
sql>CREATE TABLE table1 (a bigint);
operation successful
sql>CREATE TABLE table2 (a bigint);
operation successful
sql>SELECT (SELECT max(1) FROM table2 WHERE a=tmp.a) FROM (SELECT a FROM table1 GROUP BY a) tmp;
sql>
write error on stream

2021-09-29 16:43:03 ERR testdb[381]: mserver5: /home/niewerth/git/MonetDB/sql/server/sql_query.c:105: query_outer_used_exp: Assertion `(!is_sql_aggr(f) && sq->grouped == 0 && e->card != CARD_AGGR) || (!is_sql_aggr(f) && sq->grouped == 1 && e->card <= CARD_AGGR) || (is_sql_aggr(f) && !is_sql_farg(f) && !sq->grouped && e->card != CARD_AGGR) || (is_sql_aggr(f) && !is_sql_farg(f) && sq->grouped && e->card != CARD_AGGR) || (is_sql_aggr(f) && is_sql_farg(f) && sq->grouped && e->card <= CARD_AGGR) || (is_sql_aggr(f) && sq->grouped && e->card <= CARD_AGGR)' failed.
2021-09-29 16:43:07 MSG merovingian[34]: database 'testdb' (-1) has crashed with signal SIGABRT (dumped core)

Expected behavior
MonetDB does not crash

Screenshots
If applicable, add screenshots to help explain your problem.

Software versions

• MonetDB version number v11.39.18
• OS and version: Ubuntu 20.04
• self-compiled

[sjoerdmullender]

Assertion in query_outer_used_exp goes off.
Problem is, this is true: !is_sql_aggr(f) && sq->grouped == 0 && e->card == CARD_AGGR which is almost the first part of the assert, except for the == vs. != at the end.

[PedroTadim]

This was fixed on the default branch today and it will be available on the next feature release.