Re: icp bitrot?

From: Robert Collins <robertc@dont-contact.us>
Date: 08 Oct 2002 14:37:07 +1000

On Tue, 2002-10-08 at 14:23, Duane Wessels wrote:
> > Well I've genericalised the code in an experimental branch I've got,
> > it's down to a two if blocks, and all the sending and recieving logic is
> > 100% shared already :}. If we can have v3 do RTT lookups without
> > breaking the protocol, I can factor out even more code, giving us a
> > trivial icp_v3.c, which won't bitrot anymore (because only the
> > differences will be present in it).
>
> AFAIK, it is not safe to do RTT lookups for v3. Since the v3 protocol
> is not documented, Squid (or any non-netapp implementation) should ignore
> those two overloaded fields.

Ermm, I'm not suggesting we look at the request fields. In V2 we look at
the net db to decide if we are willing to fetch the request or not. In
v3 we don't. *Shrug*.

> I thought maybe you were digging into the code because of a
> problem or new feature. But now I understand that you are
> just bored. :-)

Actually, I'm creating async storeGetPublic* functions. This requires
breaking every synchronous function that use storeGetPublic up into a
start and completion routine. ICP uses storeGetPublic...

I tend to refactor before adding features, it makes it easier to be sure
I've done the right thing.

Rob

Received on Mon Oct 07 2002 - 22:37:12 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:16:53 MST