Dylan T.
Dylan T.
Not too sure how this will interact with #6837 currently. Since the arguments won't be `string[]` anymore.
I suppose we do also have links in every github release that tell people the correct URL to download the binaries, so this also wouldn't be a huge concern
Yeah, this is most likely to do with the prediction issues discussed in #6803. There's an unresolved discussion on how to deal with the issue.
It should be fine. We only care about byte lengths, and splittable whitespace chars will never be more than 1 byte.
@inxomnyaa raised concerns about performance when processing subcommands. I'm not sure to what extent this is a problem, but I have seen some plugins create much more complicated command signatures...
> Optimizing overload selection shouldn't be a blocker for this PR. Plugins that are particularly concerned about huge number of overloads could always fall back to RawParameter to leverage their...
I don't know about the tick stuff, that doesn't seem necessary to me, but yes there are plans to expand the system. I'm currently unsure how to deal with target...
I'm considering possible options to introduce this API in PM5 to let people adopt it early, but it's not too clear how it would work. Current thinking is to introduce...
Demo of how LuckPerms command tree might be implemented using the API in this PR as of e762b12: https://gist.github.com/dktapps/57a0a2ba161592d0e18577c4e93f6264
Yeah, target and enum are on my todo list. I've been more focused on initial system architecture so far though.