Match token but don't consume it?

[rowlesmr]

How can I describe a rule that must match a series of tokens to succeed, but leave the last rule to be consumed by a separate rule?

eg:

comment: ('#' anyprintchar* eol)+
whitespace: (" " | "\t" | eol | comment)+
singlequotedstring: "'" anyprintchar* "'" whitespace

Here, singlequotedstring must have whitespace after the final quote mark, but that whitespace is not part of the quoted string.

[erezsh]

There is no syntax in Lark for this behavior. But you can emulate it yourself if you use Lark.parse_interactive().