Perf: skip ahead when parsing URLs

https://github.com/43081j/fast-wrap-ansi/blob/4559b83e5dfa9eb3ccb5ffc5155b3b769b82e1c7/src/main.ts#L229-L231

here we know for sure the characters ahead are the URL we just matched

currently, we iterate through these one by one and append them.

maybe we can squeeze more performance out of this by immediately appending the entire URL and skipping the `i` ahead