Github user HeartSaVioR commented on the issue:
https://github.com/apache/storm/pull/1714
@arunmahadevan
I'm not sure we can have our own expression compiler which supports
features competitive to Calcite. It reminds me why this change is necessary.
We have been leaving expression compiler which has lots of unimplemented
things and also bugs. It was pushed more than 6 months before and only few of
contributors are taking care of. We have lots of milestones (not a single task)
to do even without Storm SQL, like Unified API, extending Window feature
(session, etc.), Beam integration, Java porting, Metrics, and so on. Even I
have been focusing to work on Storm SQL, I give up keeping expression compiler
because of having less eyes to watch, and error-prone way of building code
block (string concatenation). We could save amount of time if we replace this
faster, and it will save another amount of time.
Btw, normally we're often worry about dependency if dependency's activity
is low or release interval is too long. But Calcite's release interval for
minor release is even faster than Storm's bugfix release.
Calcite released 1.7.0 from Mar 22, 1.8.0 from Jun 13, and 1.9.0 from Sep
22, and 1.10.0 today. I have already fixed two bugs in unit tests and will be
shipped to 1.10.1 or 1.11.0.
(Yes in this case we need to wait on that.)
I'm not on the fence of making it pluggable, but just wondering we have
comparable plugs on that, and we can make an effort for that. Would you want to
be a sponsor on current expression compiler?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---