What is again the job of the http parser? And would it maybe make sense to have its own repository for that library, so others can use it as well without crow?
In Crow's case the parser takes the raw text sent over TCP and converts it into
crow::request objects, it would normally also parse HTTP response text but we don't care about that part.
It wouldn't really make sense to have them separate, like I said, the plan is to take something that already works and make it work with Crow's own connection and request objects (when it's intended to work with other kinds of classes/data structures)
v0.3. And there have been better ways of implementing many of its features since then. I unfortunately don't have enough time to go back and rewrite it, primarily due to focusing on Crow itself. But I would absolutely back up efforts to develop real world examples of how to use Crow for different scenarios.