[net2-wg] Re: T1->T2 porting / shim layer

[Joe Polastre]

By changing message_t, you've made life way more difficult for
porting.  The ADTs, a great idea and I wholeheartedly support, don't
lend themselves well to porting T1>T2.

Other problems you'll run into are interface name conflicts, ie Timer vs Timer2.

-Joe

On 2/8/07, Philip Levis <pal at cs.stanford.edu> wrote:
> On Feb 8, 2007, at 1:55 PM, Joe Polastre wrote:
>
> > On 2/8/07, Omprakash Gnawali <gnawali at usc.edu> wrote:
> >> The interfaces used by the components in these modules are defined,
> >> implemented, and maintained by the core WG so I think it would be
> >> something that the core WG should look into. Phil, was there ever a
> >> discussion on T1/T2 compatibility in the core WG?
> >
> > Not to beat a dead horse, but Boomerang is one possible place to look
> > to see how others have dealt with T1>T2 compatibility.
>
> Right -- the trick with Boomerang is that it's been able to leverage
> improved *implementations* but not improved *interfaces.* Its job is
> also made a lot easier by being single platform; the largest amount
> of code changes in porting often come from the fact that packets in
> T2 are ADTs, so you can't just access their fields arbitrarily.
>
> Phil
>