Indicate deprecated option aliases

[som-snytt]

Fixes #24333

Option aliases are SettingAlias which may have a Deprecation with an optional message.

[tgodzik]

Maybe it would be clearer to add a new list deprecatedAliases?

[Gedochao]

+1
that'd actually be more readable for when the API would be read outside of the compiler, as well.

[hamzaremmal]

+1

[som-snytt]

OK, the people have spoken!

[som-snytt]

rando download error

Edit: followed by need to update shapeless-3 options.

[som-snytt]

Technically, the number of parameters to a method should be capped at a gazillion, which was one reason to like the first, cute solution.

[Gedochao]

nitpick: wouldn't it increase readability to have a

case class SettingAlias(name: String, deprecation: Option[Deprecation] = None)

?
or at the very least, use a named tuple here 🤔

[som-snytt]

Thanks, I'm not accustomed to such fancy language features. I'll update it.

I found consolation in the thought that deprecated aliases are rare, and it's OK if they are awkward to write.

[tgodzik]

LGTM

[som-snytt]

Pushed @Gedochao suggestion, which improves read- and write-ability.

A unit test broke that string-interpolated alias, which changed type. I needed -Wtostring-interpolated today.

I hope the code base can progress and turn on warnings!

[som-snytt]

This will fail until shapeless-3 community project is updated.

[Gedochao]

This will fail until shapeless-3 community project is updated.

If that's the case, then I suppose this won't go in until 3.10.
cc @WojciechMazur

[som-snytt]

@Gedochao I don't understand. We can't deprecate options?

[WojciechMazur]

The PR to dotty-staging was merged, please set the submodule so that it points to current main, seems like the notification about created PR were lost.

Deprecations of existing scalacOptions would be quite important for LTS 2, so these can be possibly removed for LTS 3 if we'd need to. Base don that it's likely going to be back-ported to RC2

@Gedochao It was only a local change to replace -Xfatal-warnings with -Werror so it should be fine. It has has not made it in time for RC1 becouse it was stuck on dotty-stagging fork PR (also we'd need to re-evaluate write access to that repo, seems like currently only a single person can merge PRs there)