When running under innodb_snapshot_isolation=1, InnoDB gives error for conflicts that are not yet committed.

The following example shows this issue:
Note that if one removes the last SELECT statement or replaces that with an 'UPDATE SET b=3 where a=2', one will not get an error for the last update. Instead the last update would hang, which is what is expected.

--source include/have_innodb.inc

connect con1,localhost,root,,;

set @@innodb_snapshot_isolation=1;

connection default;

set @@innodb_snapshot_isolation=1;

create table t1 (a int primary key auto_increment, b int) engine=innodb;

insert into t1 values (1,1),(2,2);

begin;

update t1 set b=3 where a=1;

connection con1;

begin;

select * from t1;

update t1 set b=4 where a=1;

produces

mysqltest: At line 15: query 'update t1 set b=4 where a=1' failed: ER_CHECKREAD (1020): Record has changed since last read in table 't1'

Note that connection default has not done a commit or rollback yet.
What I would expect is that the last update should hang until default has done a commit or rollback.

This issue was reported by a user started to get ER_CHECKREAD errors when upgrading to 11.6.2.
I think that changing default of innodb_snapshot_isolation is the wrong thing to do for MariaDB as this will break any application that uses incremental updates (very typical in data warehouse applications) to increase performance:
UPDATE warehouse set inventory=inventory-1 where item="pc";

innodb_snapshot_isolation is a good thing to have, but probably not for the current MariaDB user base.