How to release Fable 5 stable?

[MangelMaxime]

Hello @ncave @dbrattli,

Fable 5 has been in alpha version since 1+ year now.

I think we should make a plan to release it as stable because some people won't use it / test it unless we do so.

From my point of view the main reasons for keeping Fable 5 in alpha are:

1. We introduced Nullness / Nullable support because we lack feedback on real experience with it

   For this one, I think we will only be able to gather feedback when making Fable 5 stable and documenting the feature itself.

2. We want to modify Fable.Core APIs to be more aligned with Fable multi-target nature. Because of that we don't want to say that Python for example is stable because we do have API changes planned

   I think this one would still be valuable to have done as part of Fable 5, because it would allow Python to be marked as Stable. I do have some time available this week to make a PR demonstrating what I am envisioning.

   This work will probably take a few iteration to get right, but once done it will show the path for any target.

   The most difficult one to port will be JavaScript, because we will want to keep backward compatibility for the existing libraries at least until Fable 6+.

3. We wanted to improve / rework the CLI commands to be more aligned with Fable multi-target nature.

   This is probably a must have but not mandatory, we could do it later as part of Fable 6.

4.There is the new Beam target that @dbrattli is working on which is impacting Fable.AST but that changes should be binary compatible or we could just add the new Language.Beam to the stable release of Fable.AST to prepare for it.

Do you have any other points that you see are the reason why we are keeping Fable 5 as alpha?

[dbrattli]

2. Look forwards to learn more about what you are thinking about for Fable.Core. We can also just keep Python in beta for Fable v5 if that makes things easier. My feeling is that usage for Fable.Python is still limited so its still hard to know what regressions have been made since v4. Especially for Fable.SimpleJson.Python, Thoth.Json.Python, Fable.Pyxpecto, ...

3. I don't think the Beam target will make it to Fable v5. It has 1900+ passing unit-tests but still a lot of work to be done. But we could include it as an alpha if that could draw some more attention and help to get it to a beta state. But no problem for me if this needs to wait until v6-alpha.001. Adding a new target also create a lot of extra comittment for maintainers so we should be really sure we want this. Do we know if there is any usage on targets such as Php, or Dart?