how to parse with a span offset to make parser yield SimpleSpan with correct result in multiple phases parsing?

[trim21]

I currently have many sub-parser, I want to split my file into blocks and parse block one by one. but I'm hoping to still get span of whole file so I don't need to manully fix the span offset in nexted sub parsers, is there some easy way to do this?

[zesterer]

Yes, you can use .map_with(|x, extra| ... extra.span() ...) to get the span of the input that a parser parsed. There are examples of its used in several of the repo examples, such as mini_ml.rs.

[trim21]

Thanks. I do use .map_with(|_, e|) to get span, this is where I got the problem...

My lex is not really somethings mini_ml.rs has, it doesn't "lex" input string into tokens, it just split it into multiple sub strings, each sub strings still contains multiple tokens, and the sub parsers also parse from str, not Token.

[zesterer]

Ah, I see what you mean now: my bad. Am I right that you're using the .nested_in(...) combinator?