How does varnish handle passed requests?

Question

Who can tell me how to implement (which action to return from which function) these different scenarios in varnish 3?

1. Don't look in the cache for this particular request and don't cache the response.

2. Do lookup the request but don't cache the response when missed.

In both cases the response should not but cached but later requests should be allowed to cache the response for same URL, so you can't use vcl_fetch:hit_for_pass.

Answer 1

1: vcl_recv: return pipe

2: vcl_fetch: obj.ttl = 0