bugfix: Exclude newer LSP4j, which is built on JDK 11

[tgodzik]

It seems the CI started failing because of out bump, but that was not picked up aside from windows tests weirdly.

Metals recently updated to newest LSP4j, though I think nothing new was added and no changes were done to the presentation compiler so we should be ok to downgrade locally: https://github.com/scalameta/metals/pull/6126/files

The update was mostly done so that we done fall behind in the LSP standard, but this might become a problem in the future here. Do we ever plan do drop supporting JDK 8?

[rochala]

Maybe we should consider enabling windows CI for all presentation compiler PRs to avoid situations like this ? This is the module that is most likely to be affected by different OS, and presentation compiler tests are super quick ~~ 30 sec for the test suite. @hamzaremmal

As of dropping support for JDK 8 I don't think this is happening any time soon.
But I suppose, we may be able to release presentation compiler for JDK 11+ ?

[tgodzik]

But I suppose, we may be able to release presentation compiler for JDK 11+ ?

I would opt for that if possible.

[Gedochao]

But I suppose, we may be able to release presentation compiler for JDK 11+ ?

I'll raise the topic on this week's core meeting.

[Gedochao]

Maybe we should consider enabling windows CI for all presentation compiler PRs to avoid situations like this ? This is the module that is most likely to be affected by different OS, and presentation compiler tests are super quick ~~ 30 sec for the test suite.

+1