Nitpicky question again, but I noticed the use of an ExternalTokenizer for javascript template strings (javascript/tokens.js at main · lezer-parser/javascript · GitHub) where, it seems this can be done in pure lezer as well with
templateContent {
(![\\`$] | "\\" _ | "$" ![{])+
}
InterpolationStart { "${" }
templateEnd { "`" }
Is there something I’m missing? Any gains to using an external tokenizer here?