Prepare 1.2.0 June release

[Ailrun]

Resolves partially #1890

[isovector]

[pepeiborra]

Should bump the versions of ghcide, hls-graph and his-plugin-api as well in preparation for the Hackage upload?

[Ailrun]

@pepeiborra Oh, right. I will bump those too.

[berberman]

Also, I think we should bump hls-brittany-plugin to 1.0.0.2

[Ailrun]

This changes because hlint plugin probably won't work with ghcide <= 1.3 after #1862

[hasufell]

I don't see the .tar.gz files that are usually used by ghcup.

[Ailrun]

@hasufell It's not yet built. It will be built after all other version-specific .gz files are done.

[hasufell]

the tar job seems to be failing: https://github.com/haskell/haskell-language-server/runs/2813407586?check_suite_focus=true

[Ailrun]

Yes, I'm resolving the issue.

[hasufell]

Also... since ghcup now supports windows too, we'd need a .tar.gz for windows as well.

[Ailrun]

@hasufell Ah... true. @jneira Could you generate it manually? I will add a CI entry.

[jneira]

sure, will do this night gmt

[jneira]

As usual i have to generate manually the version for 8.8.4 (#276), the only one we support for windows and it will take some time in my poor laptop

[anka-213]

Should we add the ghc-9.0 build as well to the release, so users doesn't have to build manually from source to get the basic ghc-9.0 support? Or is it to late now?

[jneira]

It is a good idea, if the executable can be used without tweaking too much configuration. A note should be added in the release, with a warning for the limited support and linking to the list of working components.
Not sure if we should change the file name to signal it is experimental and with limited support and avoid its automatic download by the vscode extension.
As stated in the new release doc version we can add new artifacts but no update old ones.

[anka-213]

No, there shouldn't be much configuration tweaking needed. Adding the flag --cabal-project=cabal-ghc901.project or copying the cabal-ghc901.project to cabal.project should be all that's needed.

I don't think disabling automatic download should be needed, as long as it is clear to the user that it is experimental, since partial support is better than no support at all.