V4 Roadmap

[timofei-iatsenko]

Let's gather here tasks for new upcoming major release:

• Extractor: Don't use intermediate catalogs refactor(extractor)!: don't use intermediate catalogs #1358
• Extractor: Change babel configuration approach, refactor(extractor/babel)!: sealed, non-fragile babel extractor options #1367 fixes:
  extract translates passed file with includes import react-native-qrcode-svg #1314
  Inconsistent extractBabelOptions handling #1305
  Typescript without React? #1206
  Cannot process file /<path>.tsx: Duplicate declaration "Trans" #952
• Macro: Update public typings refactor(macro): more accurate typings for macro + testing #1340 fixes:
  Ability to use elements within Select values #1288
  Select Macro Documentation Example is wrong #1324
  Mess in exported Macro Typescript Typings #1323
• Macro: RFC: Split macro to vanilla js + react #1361. This opens the possibility for new frameworks such as solid or qwik or vue. Also, properly fixes @lingui/macro has undeclared dependency on @lingui/react #936
• RFC: Context + hash based message id #1360 fixes:
  Message context (msgctxt) not working #1198
• deps tree Extractor. Support MPA apps (NextJS) and monorepos, feat(): Experimental Deps extractor #1469
• Custom extractors, refactor(formats): extract formats into separate packages #1536
• Provide an automatic codemod for migration v3 -> v4.

I'n going to create a separate issues (or discussion) for each of the items in the list and elaborate more on how I see these changes.

@Martin005 @andrii-bodnar what do you think about that?

[andrii-bodnar]

@thekip thank you! It's definitely great to have such a list 🚀

We can collect these issues and start working on them when we'll be ready for a major release.

Releasing a major version is a huge commitment so it should be carefully planned and should happen when it really justified.

IMO, before that, we need to resolve non-breaking issues and stabilize the library as much as possible.

[timofei-iatsenko]

You have plans "somewhere" not in public field, nobody knows them. Let's bring it to the public.

What you mean by:

we'll be ready for a major release.
"stabilize"

Actually, this list here not because I have a lot of free time, but because I want to fix existing issues and bad design to be able to move forward.
And all task in this list is about "stabilize". I don't understand what else you think should be "stabilized".

[andrii-bodnar]

stabilize

I mean just reducing the number of current issues, as much as possible keeping it backward compatible

we'll be ready for a major release

Saying that I meant when we're sure we won't have to do the next major release anytime soon after v4. Major releases are not a good thing for users, because it's always accompanied by changes in the code, support from our side, etc.

But it's ok and I understand that it should happen. Just trying to avoid extra effort for users in switching versions.

You have plans "somewhere" not in public field, nobody knows them. Let's bring it to the public.

The main plans for now:

• reducing the number of current issues as much as possible
• new documentation
• export to a new file format (this idea is still in draft and needs more investigation)

[timofei-iatsenko]

reducing the number of current issues as much as possible

To fix most of the issues which appear every day, we need to introduce breaking changes. They would not be noticeable for most of the users but it's still breaking change which should bump a major version.

[Bertg]

A while ago I suggested in #1283 the following:

Lingui (in production) produces code with translations baked in, basically a package per language, with 0 runtime impact.

Could this still be considered for v4? I understand it's a radical change, and probably requires a very different set of skills. I'm just wondering. If not would this be v5, or a "never"?

Could an other package build on top of v4 to make this possible, to opt-in to it?

Also, wan tot say, I'm very happy to see activity here again. I'm very grateful for the work you guys are doing!

[timofei-iatsenko]

@Bertg see #1384

But it's not planned for v4 for now.

[andrii-bodnar]

@Bertg thanks for your suggestion! The v4 roadmap already includes a lot of tasks so we're not sure if #1384 will be included in this release. But definitely, it's not a "never" perspective 🙂

[andrii-bodnar]

Created a milestone for related issues tracking - https://github.com/lingui/js-lingui/milestone/7

[andrii-bodnar]

The Pre-Release is here 🚀

v4.0.0-next.0