@MIchaelMainer @zengin the older pipeline seems to still be active given the number in this branch. Did we forget to disable something?

@baywet this is the result of my full run on main.

But should we have a difference given we already merged #655 ? Or was the metadata updated in between the two?

That one was running in @MIchaelMainer's own ADO instance (small build number: 50). It is hard to tell for me what particular code that ADO instance is running. The correct end result should be this one.

That explains a lot! The other one was only a comments change. I'll revert minor bump and bump patch instead on this one tomorrow and we'll call it a day.

@zengin Because this update added a lot of changes, using the typesummary was fastidious.
I used pkgdiff instead, it works well for java on WSL. Due to how the diff are structured, we could imagine parsing the outcome files and automate the regression check. What do you think?
github.zip

@baywet, I like it. It looks great.