[#5647] improvement(build): update the jar name to add gravitino prefix

[orenccl]

What changes were proposed in this pull request?

As title

Why are the changes needed?

Fix: #5647

Does this PR introduce any user-facing change?

Makes users easy to know it is from Gravitino

How was this patch tested?

Execute Gradle command ./gradlew publishToMavenLocal -x test then check in Maven local path.

[orenccl]

I also encountered this network error on other pull requests. Is there a way to improve this?

> Task :clients:client-python:downloadHadoopPack FAILED
* What went wrong:
Execution failed for task ':clients:client-python:downloadHadoopPack'.
> A failure occurred while executing de.undercouch.gradle.tasks.download.internal.DefaultWorkerExecutorHelper$DefaultWorkAction
   > java.net.SocketException: Network is unreachable

It seems to be a network issue in the CI environment.

[jerryshao]

It is occasional, maybe you can configure that gradle plugin to increase the timeout or add retry times (but it depends on if that plugin has such configurations).

[orenccl]

I've discovered that the artifactId is the jarName without the version.
(Reference: Maven Naming Conventions).

Additionally, the artifactId defaults to project.name.
(Reference: Gradle Documentation for MavenPublication).

Therefore, setting the artifactId only once in the root build.gradle.kts should resolve the issue.

[jerryshao]

For some modules like spark, flink, we already added "gravitino-" as the project name, can you please check if your change will add the prefix again, also please align these modules with other modules.

[orenccl]

After looking into it, I found that the artifactId set in the root project doesn’t apply if the child build.gradle.kts files have their own artifactId settings.

Moreover, it also looks like they need to do some custom work to get their naming conventions right.

For example

val scalaVersion: String = project.properties["scalaVersion"] as? String ?: extra["defaultScalaVersion"].toString()
val sparkVersion: String = libs.versions.spark33.get()
val sparkMajorVersion: String = sparkVersion.substringBeforeLast(".")
val baseName = "${rootProject.name}-spark-connector-runtime-${sparkMajorVersion}_$scalaVersion"

[tengqm]

Why the invocation to Locale.getDefault()?

[orenccl]

I was following the code style in this script, where toLowerCase(Locale.getDefault()) is actually just toLowerCase().

[jerryshao]

Specifying Local for lowercase () API is a best practice in JVM language, otherwise we will get an expected result.

[orenccl]

Specify locale to en_us
specify Locale

[jerryshao]

I tested locally, LGTM, @orenccl do you have anything else needs to be addressed?

[orenccl]

Hi, @jerryshao
I believe this is good to go, but happy to address any further comments.

[jerryshao]

@tengqm do you want to take another look?

[tengqm]

@tengqm do you want to take another look?

Good to learn this convention.
LGTM

[tengqm]

@tengqm do you want to take another look?

Good to learn this convention. LGTM

p.s. I raised the question because I was checking the use case here. It is not about converting a general text message which are supposed to be localized. My understanding was that JAR names are always ASCII printable characters and we don't need (want to) over-engineer this to consider file names with "イ" or "い".

[jerryshao]

@tengqm do you want to take another look?

Good to learn this convention. LGTM

p.s. I raised the question because I was checking the use case here. It is not about converting a general text message which are supposed to be localized. My understanding was that JAR names are always ASCII printable characters and we don't need (want to) over-engineer this to consider file names with "イ" or "い".

I see. Yes, it depends. But some static check tools like error-prone, or IDEA itself may complain about this if the Locale is not set. So typically, we will add the Locale to the methods like this.

[jerryshao]

Hi @tengqm I'm going to merge this if there's no further comment.

[tengqm]

no objection from me