[
https://issues.apache.org/jira/browse/FLINK-13414?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17534962#comment-17534962
]
Roman Grebennikov edited comment on FLINK-13414 at 5/12/22 11:20 AM:
---------------------------------------------------------------------
I went further using [~NickBurkard] 's flink-adt scala3 PR, and cross-built a
forked version of Flink's Scala API:
[https://github.com/findify/flink-scala-api]
Check the github repo for details, but if you can afford re-bootstrapping your
job state from scratch, then migration is quite straightforward. Main features:
* Cross-built for Scala 2.12, 2.13 and 3.x
* 100% replica of the 1.15 Scala API, with deprecated methods removed.
* ClosureCleaner from Spark 3.x with 2.13 and java 17 support.
* Agnostic to the serialization format, but can use flink-adt to support scala
more natively.
It's still an experimental version which is not yet well tested, but I don't
see any possible dangers in it:
* we're using flink-adt exclusively for a lot of Flink jobs in Findify for
almost a year with no issues so far.
* Flink's Scala API is a thin wrapper on top of Java API, so there is nothing
to break.
The only not well-tested thing is scala3 derivation, which is a compile-time
process. So if you have compilation errors due to flink-adt failing for your
case classes, you're welcome to submit a bug report on github.
was (Author: rgrebennikov):
I went further using [~NickBurkard] 's flink-adt scala3 PR, and cross-built a
forked version of Flink's Scala API:
[https://github.com/findify/flink-scala-api|https://github.com/findify/flink-scala-api.]
Check the github repo for details, but if you can afford re-bootstrapping your
job state from scratch, then migration is quite straightforward. Main features:
* Cross-built for Scala 2.12, 2.13 and 3.x
* 100% replica of the 1.15 Scala API, with deprecated methods removed.
* ClosureCleaner from Spark 3.x with 2.13 and java 17 support.
* Agnostic to the serialization format, but can use flink-adt to support scala
more natively.
It's still an experimental version which is not yet well tested, but I don't
see any possible dangers in it:
* we're using flink-adt exclusively for a lot of Flink jobs in Findify for
almost a year with no issues so far.
* Flink's Scala API is a thin wrapper on top of Java API, so there is nothing
to break.
The only not well-tested thing is scala3 derivation, which is a compile-time
process. So if you have compilation errors due to flink-adt failing for your
case classes, you're welcome to submit a bug report on github.
> Add support for Scala 2.13
> --------------------------
>
> Key: FLINK-13414
> URL: https://issues.apache.org/jira/browse/FLINK-13414
> Project: Flink
> Issue Type: New Feature
> Components: API / Scala
> Reporter: Chaoran Yu
> Priority: Minor
> Labels: auto-deprioritized-major, pull-request-available
>
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
