lindera
lindera copied to clipboard
About implementing the Clone trait
Tokenizer/Segmenter/Dictionary supports Clone, but it's very expensive when dictionary data is the Vec variant, is there any practical use case for clone? if so, we should probably use Rc/Arc to reduce the cost, otherwise I think we should just remove the clone code.
(this is separated from https://github.com/lindera/lindera/issues/443 for tracking)