[mythtv] Protocol negotiation?

[David Shay]

Michael Tiller wrote:

>
>
> On 9/29/06, *Geoffrey Kruse* <gkruse at gmail.com 
> <mailto:gkruse at gmail.com>> wrote:
>
>
>     Well, I wouldn't expect to support any protocols prior to the one
>     where this is implemented.  (Same as the current situation)
>
>
> Yes, I had that in mind too.  But given the protocol description, 
> wouldn't it be trivial to make a proxy that provided a "MythTV 
> protocol XYZ" compatible socket.  Do you see what I'm getting at?  If 
> the backend is really on port 6398 (I don't know the real port number 
> right now), couldn't you create a proxy at say, 10030 and have it act 
> as a simple proxy for 6398 but just rewriting the string you describe? 
>   Then you could just point an old client to the proxy.  Are am I 
> missing something?
>
Yes.  The real difficulty is dealing with the logic, not just string 
constants.  Look, for example, at the livetv changes -- these were new 
commands, changed commands, etc.

mvpmc does provide backwards compatibility, at least for most things, 
but that is at the cost of complex code that checks the version and does 
different things.  I agree with an earlier poster that there is no point 
in building this into mythtv -- it's up to other connecting apps to deal 
with the backwards compatibility if they feel that it is necessary.  
Also, the upnp stuff in .20 provides a compatible/version-neutral way 
for 3rd-party apps to integrate with myth for at least some things.