GHA JAR comparison correctness

Hello!

• Vote on this issue by adding a 👍 reaction
• If you want to implement this feature, comment to let us know (we'll work with you on design, scheduling, etc.)

Issue details

Enhance GHA to define the proper old JAR we want to benchmark against and make sure it's getting the proper old JAR as publishing matures.

Comments by @t0yv0 on

• #327

This is really worrisome to me that it's inline with the main build and YAML quickly grows. What's worrisome is that one day we will accidentally build the wrong jar and publish the wrong jar or else test the wrong jar.

Would it be possible to checkout the old version and perhaps separate this size check a little more?

I guess we're fine for now but we can revisit later on if we refactor/grow this build file.

One possibility is that we have SNAPSHOT builds published from main into the Maven repo. Those can be a point of comparison without rebuilding.

Affected area/feature

Start by inspecting the repository's GitHub Actions workflow and the context in #327. Determine how the old JAR is selected for benchmarking, then define and verify a stable comparison against the intended published artifact without relying on an accidentally wrong build output.