Add `dangerously_pass_through_form_requests` to support extractor-based classical forms?

[LeoniePhiline]

Library users will be responsible for making sure CSRF checks are performed from form data. (Consuming the body is necessary for classic, formdata-based CSRF protection.)

[LeoniePhiline]

Much better:
Rather implement a second middleware which actually intercepts form requests (digesting the request body for form-Content-Type non-safe request methods),
and injects hidden token fields into all forms found in the response HTML.

This causes some indirection and might not be the very fastest, but the most secure way.