feat: resolve code action

[sigmaSd]

fix #5118

tested with deno (works)

I imagine the code is not ideal since I'm blocking for a future inside the callback to wait for the resolver also I had to clone the code action

[pascalkuthe]

instead of blocking on the future you can/should add new job that applies the edit (and the command)

You also need to add support for resolving commands, and set the capabilities correctly. While you are it, it would be good to add some capabilities we are missing. RA tends to frreqendly ignore capabilities so that meant we forgot to add them sometimes. Simply add the following to the code action capabilities

                        is_preferred_support: Some(true),
                        disabled_support: Some(true),
                        data_support: Some(true),
                        resolve_support: Some(CodeActionCapabilityResolveSupport {
                            properties: vec!["edit".to_owned(), "command".to_owned()],
                        }),

[sigmaSd]

As far as I can tell, I can't nest jobs callback, I searched the code base and there is no other instance of nesting callbacks

There seems to be a similar lsp resolve call, but its registered to be triggered with idle timeout https://github.com/helix-editor/helix/blob/master/helix-term/src/ui/editor.rs#L1010 (which require saving state in the editor)

[sigmaSd]

Another similar thing is this https://github.com/helix-editor/helix/blob/master/helix-term/src/commands.rs#L4513

set_completion does block internally (it uses tokio_blockon internally in Completion::new to resolve items)

[pascalkuthe]

hmm right I forgot this was already inside a box, I have a patch locally to allow that but its a bit more involved so that's probably ok for now

[pascalkuthe]

you should still try to resolve the command tough in case its none