Library publicity

[mrlem]

It took me some time to stumble on Java GI (I did so indirectly), don't you think the library could have its entry in GTK bindings list now?

I found this 2022 gnome discourse discussion where you mentioned you wanted the lib to be more mature before this happens, and was wondering what your thoughts were now. I personally think the binding is very decent and completely apt to produce nice apps, and it would gain from a broader audience, which this list could bring (as that's the natural place to check for a way to create gnome apps using Java).

[jwharm]

Thank you for your interest in java-gi and for the kind words!

I know about the GTK website.
And there's a similar page on the GNOME website.

Personally I want to finish a few more backlog items before I consider the library stable enough to recommend it more widely:

• Now that I registered java-gi.org, I want to move everything from io.github.jwharm.javagi to org.javagi (maven coordinates, module and package names) but this is a breaking change for existing users, so I want to first publish a smaller 0.12.3 release with bugfixes and small enhancements, and only use the new domain from 0.13.0 onwards.
• I want to implement a decent regression test suite, using gobject-introspection-tests. I started working on this a little while ago.
• I'd like to use JSpecify nullability annotations instead of Jetbrains Annotations. This might be a breaking change for some users (I'm not sure yet).
• I want to extend the "HelloTemplate" example, to be a complete Flatpak application with icons, translations, resources, settings, a .desktop and a metainfo.xml file. This is relatively easy to do with a Meson build script. (Meson would still execute Gradle as a custom build step to compile the application.)

Now, this list used to be a lot longer 😄 so it's probably just a few months more, before I'd say it's "good enough". But it's a weekend/evening project, so I can't make any promises. If there's anything you'd like to contribute, let me know!

[mrlem]

Thanks for the insights, all those points seem indeed valuable.
I'll try to help. For instance, I can look into the JSpecify annotations migration (and at least see whether it causes errors / warnings in the existing examples & my own pet project).

[jwharm]

That'd be appreciated!

The nullability annotations are currently only set on parameters, not return values. And array elements and generic types aren't annotated yet. The current code is very simple:

• CallableGenerator.generateMethodParameters
• Parameter.nullable/notNull

With JSpecify, we can annotate the entire class @NullMarked (which means everything is default not-null), and add @Nullable where applicable.

[jwharm]

@mrlem I have a rather large PR (#238) ready to merge into main, changing the packages and modules from io.github.jwharm.javagi to org.javagi. If you're working on a PR for JSpecify annotations, I can wait with merging until you're ready? Because #238 moves all source files around and will make merging other changes a hassle.

[mrlem]

@jwharm don't wait for me, go for it. I'll rebase what I have done (didn't have as much time as I wanted this week, and won't have much more time this week-end)

[jwharm]

Hey @mrlem are you still looking into the JSpecify nullability annotations? Otherwise I'll probably give it a go by myself.

[mrlem]

No, sorry, I got caught in offline-life stuff. Feel free to go for it.