[SuperMuel]

Updates the example workflows and the README:

• Bump actions/checkout, setup-python, setup-node and setup-go to v7, and moonrepo/setup-rust to v1. Their breaking changes do not affect these examples.
• Use Python 3.13 instead of 3.9, which is no longer supported.
• Use the current macro runner label in the Go example. The old codspeed-macro label still works.
• Fix the README Rust workflow: name was in the wrong place, so GitHub could not read the file.
• Add examples for Java, Kotlin and Scala (JMH with Gradle) and for CLI commands (codspeed.yml). Now examples/ covers every language of the Supported Languages page. The new examples follow the workflows of the docs pages, with actions/setup-java v6.

I ran the 9 examples on small test projects, on ubuntu-latest because the test organization has no macro runner. Every job passed and uploaded its results to CodSpeed. On ubuntu-latest, the 3 JVM examples needed ./gradlew jmh --no-daemon: with the Gradle daemon, the CodSpeed step does not end after the build. The macro runner does not have this problem: the codspeed-jvm CI runs ./gradlew jmh there.

Closes COD-3885

[adriencaccia]

Let's add new example of the missing languages that we support in https://codspeed.io/docs/benchmarks/overview
Otherwise lgtm

[SuperMuel]

Let's add new example of the missing languages that we support in https://codspeed.io/docs/benchmarks/overview

I added the 4 missing languages in 1b1a3d1

And I apologize for the meat-proxying, here's Claude explanation of a problem and questions

---

Problem: on ubuntu-latest, the 3 JVM examples never finish. ./gradlew jmh builds and runs the benchmarks in about 1 minute, then the CodSpeed step waits until the job times out. This happened in 4 out of 4 runs.

Cause: in walltime mode, the CodSpeed step waits until every process that the command started has exited. Gradle starts a daemon that keeps running after the build: by default, Gradle stops an idle daemon only after 3 hours. Control runs with runner 5.6.0:

• ./gradlew jmh --no-daemon, or ./gradlew jmh && ./gradlew --stop: the step ends in about 2 minutes.
• sh -c 'sleep 240 &': the step waits the full 4 minutes. The result is the same when sleep has its output closed, and with CODSPEED_ISOLATION=false.

The codspeed-jvm CI runs ./gradlew jmh on a macro runner and does not hang. So I kept the examples the same as the docs: macro runner and ./gradlew jmh.

Questions:

1. Is this expected? Should the runner stop waiting for background processes when the command exits?
2. Until then, should we add --no-daemon to the 3 examples and to the JVM docs pages, so that they also work on GitHub-hosted runners?