aglinxinyuan commented on code in PR #6225: URL: https://github.com/apache/texera/pull/6225#discussion_r3541403629
########## amber/build.sbt: ########## @@ -276,7 +276,7 @@ libraryDependencies += "com.konghq" % "unirest-java" % "3.14.5" libraryDependencies += "com.github.marianobarrios" % "lbmq" % "0.6.0" // https://mvnrepository.com/artifact/org.jooq/jooq -libraryDependencies += "org.jooq" % "jooq" % "3.14.16" +libraryDependencies += "org.jooq" % "jooq" % "3.21.6" Review Comment: Good catch — addressed in the latest push. Instead of 3.21.6 everything is now aligned at **3.19.36**: `amber`, `computing-unit-managing-service`, `common/dao`, and the `jooq-codegen` plugin all move together, so there's no cross-module skew. The version isn't 3.21.6 because jOOQ 3.20+ ships Java 21 bytecode and we build/run on Java 17 — that's what broke CI here (`UnsupportedClassVersionError`); 3.19.36 is the last Java-17 release. `jooq-codegen` had to move in lockstep too: with a 3.16 codegen against a 3.19 runtime, the generated `TableImpl(Table, ForeignKey, Table)` call is deprecated and `-Werror` fails `DAO / Compile`. The 42.7.4 pgjdbc pin stays since the KEY_SEQ codegen fix only lands in jOOQ 3.20+. I kept the four literals rather than a shared `jooqVersion` val to keep the diff focused — happy to centralize if you'd prefer. See #6248 for the Java 21 move that would let us go past 3.19. ########## computing-unit-managing-service/build.sbt: ########## @@ -47,7 +47,7 @@ libraryDependencies ++= Seq( "io.dropwizard" % "dropwizard-core" % dropwizardVersion, "io.dropwizard" % "dropwizard-auth" % dropwizardVersion, // Dropwizard Authentication module "io.kubernetes" % "client-java" % "21.0.0", - "org.jooq" % "jooq" % "3.14.16", + "org.jooq" % "jooq" % "3.21.6", Review Comment: Resolved in the latest push — `amber`, `computing-unit-managing-service`, `common/dao`, and `jooq-codegen` are all aligned at **3.19.36** now, so there's no cross-subproject version skew. Capped at 3.19.36 (the last Java-17 jOOQ release) rather than 3.21.6, which is Java 21 bytecode and fails at runtime on our Java 17 matrix. Details in the PR description; #6248 tracks the Java 21 move that would let us go further. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
