[
https://issues.apache.org/jira/browse/FLINK-34193?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17810428#comment-17810428
]
Martijn Visser commented on FLINK-34193:
----------------------------------------
That's one way to deal with it yes, but one thing that might happen is that you
will have clashes with the Jackson or SnakeYAML version that a user might
provide. In case there's an alternative for Jackson/SnakeYAML in the code base
(like a plain Java implementation), that could be desirable over your proposed
solution.
> Remove usage of Flink-Shaded Jackson and Snakeyaml in flink-connector-kafka
> ---------------------------------------------------------------------------
>
> Key: FLINK-34193
> URL: https://issues.apache.org/jira/browse/FLINK-34193
> Project: Flink
> Issue Type: Technical Debt
> Components: Connectors / Kafka
> Reporter: Martijn Visser
> Priority: Critical
>
> The Flink Kafka connector doesn't have a direct dependency in the POM on
> flink-shaded, but it still uses the shaded versions of Jackson and SnakeYAML
> in {{YamlFileMetaDataService.java}} and
> {{KafkaRecordDeserializationSchemaTest}}
> Those cause problems when trying to compile the Flink Kafka connector for
> Flink 1.19, since these dependencies have been updated in there. Since
> connectors shouldn't rely on Flink-Shaded, we should refactor these
> implementations
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
