[jira] [Commented] (SPARK-32702) Update MiMa plugin

["Sean R. Owen (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-32702?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17184708#comment-17184708 ] 

Sean R. Owen commented on SPARK-32702:
--------------------------------------

The "not analyzing binary compatibility" messages look as expected.

So I think MiMa is stricter now or something?
Presumably, hopefully, anything the old version didn't catch isn't a real issue.

Hm, one strange thing here is that it seems to be comparing vs 2.4.0. MimaBuild.scala should at least be comparing vs 3.0.0 at this point in master. Maybe that helps reduce the change its looking at, thus the new additional errors, so I think that's valid to fix.

I think we'd have to review the new warnings and double-check them, yeah, but, I'd assume they're false positives. Therefore it'd be OK to suppress them in a PR and take a look. But is this saying there are thousands more? wow. We may need to bulk turn off some rules or something.

PS I think it would also be OK to remove the sections of MiMaExcludes.scala pertaining to, say, Spark 2. Those aren't relevant in the build at master, and, may accidentally retain an exclusion that is no longer valid, covering up an actual change.

> Update MiMa plugin
> ------------------
>
>                 Key: SPARK-32702
>                 URL: https://issues.apache.org/jira/browse/SPARK-32702
>             Project: Spark
>          Issue Type: Sub-task
>          Components: Build
>    Affects Versions: 3.1.0
>            Reporter: Denis Pyshev
>            Priority: Minor
>         Attachments: core.txt, graphx.txt, mllib-local.txt, mllib.txt, sql.txt, streaming-kafka-0-10.txt, streaming.txt
>
>
> As a part of upgrade to SBT 1.x it was found that MiMa (in form of sbt-mima plugin) needs to be upgraded too.
> After changes in build file to apply plugin changes between versions 0.3.0 and 0.7.0 binary incompatiblity errors are found during `mimaReportBinaryIssues` task run.
> In summary there are of two types:
> unclear, if should be checked or excluded
> {code:java}
> [error] (examples / mimaReportBinaryIssues) mimaPreviousArtifacts not set, not analyzing binary compatibility 
> [error] (assembly / mimaReportBinaryIssues) mimaPreviousArtifacts not set, not analyzing binary compatibility 
> [error] (tools / mimaReportBinaryIssues) mimaPreviousArtifacts not set, not analyzing binary compatibility 
> [error] (streaming-kafka-0-10-assembly / mimaReportBinaryIssues) mimaPreviousArtifacts not set, not analyzing binary compatibility
> {code}
> and found compatibility errors
> {code:java}
> [error] (mllib-local / mimaReportBinaryIssues) Failed binary compatibility check against org.apache.spark:spark-mllib-local_2.12:2.4.0! Found 2 potential problems (filtered 2) 
> [error] (graphx / mimaReportBinaryIssues) Failed binary compatibility check against org.apache.spark:spark-graphx_2.12:2.4.0! Found 3 potential problems (filtered 3) 
> [error] (streaming-kafka-0-10 / mimaReportBinaryIssues) Failed binary compatibility check against org.apache.spark:spark-streaming-kafka-0-10_2.12:2.4.0! Found 6 potential problems (filtered 1) 
> [error] (sql / mimaReportBinaryIssues) Failed binary compatibility check against org.apache.spark:spark-sql_2.12:2.4.0! Found 24 potential problems (filtered 1451) 
> [error] (streaming / mimaReportBinaryIssues) Failed binary compatibility check against org.apache.spark:spark-streaming_2.12:2.4.0! Found 124 potential problems (filtered 4) 
> [error] (mllib / mimaReportBinaryIssues) Failed binary compatibility check against org.apache.spark:spark-mllib_2.12:2.4.0! Found 636 potential problems (filtered 579) 
> [error] (core / mimaReportBinaryIssues) Failed binary compatibility check against org.apache.spark:spark-core_2.12:2.4.0! Found 1030 potential problems (filtered 1047)
> {code}
> I could not take a decision on my own in those cases, so help is needed.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org