[Moshi 2] Release alpha.1 missing from Maven Central

[dniHze]

Not really a bug. As far as I can see, the first alpha release for Moshi 2 was cut, but the publishing pipeline has failed, although the logs are hidden. Just checking if that was intended, and for other release tag lurkers out there.

[hbb20]

Any update on this? Stuck with reflection for custom annotations. without this.

[SteinerOk]

Hi @JakeWharton, @ZacSweers, @swankjesse,

I completely understand that this is an open-source project maintained in your free time, so I have absolutely no expectations or demands regarding the library.

I am simply reaching out to better understand how to plan my own strategy for using it. What are your plans for Moshi? Do you intend to continue its development, or would you recommend migrating to other JSON parsing libraries? Overall, how do you see the future of Moshi?

Thank you in advance for your time!

[JakeWharton]

I will speak only for myself here.

Moshi was the most tastefully designed JSON library for Java, and that design translates poorly to Kotlin. Now that I don't write Java, I don't use Moshi. This makes me sad. I also don't recommend Moshi to others writing Kotlin. I recommend kotlinx.serialization because while it has many things I think are worse it has enough that's better to justify its use.

If it were solely up to me (which it is not), I would revert the library to be written in Java, deprecate the built-in Kotlin support, and tell people who are currently using it to use MoshiX and/or migrate away. Then I would leave it alone. It's a great candidate for a library that can just be mostly "done".

I have no interest in maintaining it going forward. I only keep up on the Renovate PRs because it's easy and I have a somewhat misguided sense of pride because the library is named after my dog.

[dniHze]

K2 break-in issues just helped my team to commit to kotlinx.serialization transition faster than otherwise. We considered the MoshiX route, yet kotlinx.serialization was on our radar for quite a long time since we eliminated almost entirety of Java from our codebase. Not every team or codebase can afford it, in all fairness, but I agree this is better way forward.

Moshi is still a brilliant tool, and l don't want to downplay the importantce of Moshi in my last dozen of projects.