[
https://issues.apache.org/jira/browse/FLINK-30927?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17684726#comment-17684726
]
Krzysztof Chmielewski commented on FLINK-30927:
-----------------------------------------------
Provided PR above is fixing reported issue.
However CI build was not failing due to this problem. The reason why it was not
failing is that code splitter has a safety net, that whenever rewritten code
fails the compilation, Flink tries to use original code + print failing class
into the logs. That is how the problem was spotted.
Maybe it would worth to add an enhancement such this issue would in fact failed
the build? A separate issue?
> Several tests started generate output with two non-abstract methods have the
> same parameter types, declaring type and return type
> ----------------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-30927
> URL: https://issues.apache.org/jira/browse/FLINK-30927
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Reporter: Sergey Nuyanzin
> Assignee: Krzysztof Chmielewski
> Priority: Major
> Labels: pull-request-available
>
> e.g.
> org.apache.flink.table.planner.runtime.stream.sql.MatchRecognizeITCase#testUserDefinedFunctions
>
> it seems during code splitter it starts generating some methods with same
> signature
>
> {noformat}
> org.codehaus.janino.InternalCompilerException: Compiling
> "MatchRecognizePatternProcessFunction$77": Two non-abstract methods "default
> void MatchRecognizePatternProcessFunction$77.processMatch_0(java.util.Map,
> org.apache.flink.cep.functions.PatternProcessFunction$Context,
> org.apache.flink.util.Collector) throws java.lang.Exception" have the same
> parameter types, declaring type and return type
> {noformat}
>
> Probably could be a side effect of
> https://issues.apache.org/jira/browse/FLINK-27246
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
