Want cooked hs desc parse-and-decode

IRC conversation about what to do with the result of HsDesc::parse. We should do as discussed.

18:18 <+nickm> Diziet: I maintain that SignatureGated is an okay type to get 
               back from HsDesc::parse.  It takes a blinded_onion_id to make 
               sure that you got the descriptor that you want, and returns a 
               SignatureGated.
18:18 <+nickm> That SignatureGated has all the keys it needs to check the 
               signatures; you don't need to add any more
18:18 <+Diziet> nickm: So, re validation, the thing I'm missing is a function 
                that takes the output of HsDesc::parse plus all necessary keys 
                (client secrets, expected HsId and HsBlindId) and gives you the 
                inner content to use.
18:19 <+Diziet> Oh, so it promises that if you get Ok all you have to do is 
                check some self-signature ?
18:19 <+Diziet> Isn't there more than one signature in this thing ?
18:19 <+nickm> Right.  So you call   
               HsDesc::parse(s).check_signature()?.check_valid_at(&now)? and 
               then you have an EncryptedHsDesc.  
18:20 <+nickm> then you call _decrypt_ on that, and you need to check _its_ 
               inner signatures and time-bounds, since they are on the inner 
               part too
18:20 <+Diziet> Which you then call decrypt on.  And then .check_signature and 
                ... 
18:20 <+nickm> yup
18:21 <+nickm> the parse_desc_good() test goes all the way from a string to a 
               decrypted HsDesc.  It uses unwrap, but other than that the logic 
               is right
18:21 <+Diziet> OK, I wasn't confident from the doc comments and API types that 
                that would be correct.  (Indeed, you have to quite follow some 
                breadcrumbs to even find it.)
18:21 <+nickm> Feel free to add more doc comments?
18:22 <+Diziet> nickm: OK how about we invent a function that does all these 
                things ? :-)
18:22 <+Diziet> And also some doc comments yes.
18:22 <+nickm> Like, HsDesc::parse_decrypt_and_check() ?  That would be ok.
18:22 <+Diziet> Yes, that.
18:22 <+Diziet> It would give you a TimerangeBound.
18:23 <+Diziet> Perhaps several, since maybe you want bits from the middle or 
                outer too.
18:23 <+Diziet> Oh HsDesc is already that
18:23 <+nickm> yup
18:23 <+Diziet> OK great.
18:24 <+Diziet> I will make a ticket containing this conversation.
18:24 <+Diziet> Thanks for explaining