Additional strategy based on SQL code and effective data changes for simplifying mysqltest scripts
Our current experience shows that there are tests where automatic simplification is extreme slow.
The simplifier somehow picks unfortunate slices of the test for removal attempt.
There are also tests with some strict logic (example: replication tests which avoid the generation of
statements which are unsafe in some binlogging mode). Such a logic is unknown to the simplifier
and therefore it is to be expected that he is slow or produces garbage.
It would be nice to have an additional simplifier strategy based on the following fact.
It is very likely that
- a statement which fails
- a INSERT/
could be removed without changing the effect we are searching for.
Blueprint information
- Status:
- Not started
- Approver:
- None
- Priority:
- Undefined
- Drafter:
- None
- Direction:
- Needs approval
- Assignee:
- Random Query Generator Team
- Definition:
- New
- Series goal:
- None
- Implementation:
-
Unknown
- Milestone target:
- None
- Started by
- Completed by
Related branches
Related bugs
Sprints
Whiteboard
The following has been implemented:
- the algorithm has been replaced with a much faster one
- individual MTR executions have been sped up by using appropriate MTR options
- headers and footers are protected from simplification by specifying them in a separate config file
- simplification attempts where replication-unsafe statements are generated will be marked as failing
