So, when I do get back into my regular development cycle, I’m going to be working on 1.0 support. Yay 🙄

Breaking Changes: https://join-lemmy.org/news/2025-02-03_-_Breaking_Changes_in_Lemmy_1.0

Just going to get my gripes out of the way first: The Lemmy devs have no fucking clue how to version an API. They added an API v4 with some legitimately welcome changes. That’s great. But the whole reason you version your API is to not break previous API versions. Random shit is breaking in v3 on 1.0.0-alpha because they changed the v3 behavior instead of implementing all the breaking changes in the new API v4.

I tested Tess 1.4.31 against voyager (that’s as far as I ever want to touch .ml ugh), and there’s a lot of stuff that needs to be fixed. Maintaining limited compatibility with older API versions (specifically the remote community browser) is also going to be a thing because of the moronic breaking changes that shouldn’t be breaking if the devs had any goddamned sense.

Have I mentioned how much I hate the Lemmy devs?

The Answer is 'a lot'

The answer is “a lot”

I’m not sure how far out 1.0 is, but don’t expect Tess to be ready on launch day. That said, 1.0 support is going to be developed in a new version (1.5 series) that bumps the minimum API to 1.0 because, again, the Lemmy devs have no fucking clue how to do versioned APIs and only concern themselves with Lemmy UI and everyone else be damned.

I’m so tired, y’all.

  • jetA
    link
    fedilink
    English
    arrow-up
    1
    ·
    2 days ago

    I’ve always seen the first iteration of a federated protocol as a poc, to attract sensible developers who understand that API is a contract