what project structure could I use to test a version catalog that I'm creating? not in toml format, one to publish. I want to test that the accessors are created properly and that I can actually download the jars. (note: currently having a laugh as an AI bumbles through trying to make a second gradle project...)

That's it I'd say, yes. Another project that uses it. Can be a standalone test project, or you can use TestKit. Just like with any other build logic functional tests.

Yeah I suppose one could use testkit. I hate testkit so much. If I use a composite build though I don't know how it could depend on the output of the generate toml task or whatever other outputs might get published. Unless I simply manipulated that as ci build thing which means it wouldn't be a true composite build

Oh, right, consuming the version catalog from composite build does not work yet: https://github.com/gradle/gradle/issues/18847 So you would need to publish to some local directory repository and consume it from there for the test.

Thanks for the link, another example of Gradle not finishing a feature for years and years and years

You can only finish as many features as you have developers and contributors. 🤷‍♂️

Yeah, but I kind of feel like they keep adding features without finishing other ones... And not like simple features like super complicated things I'm arguably talking about this new DSL which I totally think is the wrong direction. But we'll see at some point soon I need to try that see if it works entirely different subject we don't really need to get into opinions on how Gradle does features but there were features that are even older than the version catalog that I feel like are unfinished

Yes 😞

@Oleg Nenashev this

I cannot say much about the feature itself. It is in the backlog of the JVM team, so CCing @Sterling

example of Gradle not finishing a feature for years and years and years

You can only finish as many features as you have developers and contributors.

And, as you can never have enough developers, investing in a community, contributor experience, and ecosystem growth is what many companies do 🙂 I am looking into some initiatives, but I have to admit that the Gradle core is not a priority for me as of now. I am not sure about this particular change and whether it can be realistically done by an external contributor.