PDA

View Full Version : How does "Manually add a server" work for HME?


wmcbrine
01-20-2008, 03:43 PM
So how does the "Manually add a server" function work for HME apps, anyway? Put another way, how does apps.tv work? And before anyone answers in this vein, I don't mean how to use it; I mean, how is it implemented?

I know how this works for HMO (and I implemented it for pyTivo): the client connects to port 2190, and client and server exchange Tivo Connect Discovery packets, just like the HMO server broadcasts to the local subnet. But HME doesn't use TCD; it uses ZeroConf. And it doesn't listen on port 2190... at least the sample hosting environment doesn't, and port 2190 doesn't seem to be open at apps.tv either. But when I tried the manual connection to a machine here that was running no Tivo services, the only connection attempt I saw was to port 2190. Does the Tivo act differently when attempting to connect to a host outside the local subnet? Short of replacing my switch with a hub so I can packet sniff it, I'm stumped.

apps-tv
01-20-2008, 08:25 PM
It connects to port 80 and requests a file called TiVoConnect

http://209.97.196.52/TiVoConnect

which needs to return an xml file (described in the HMO docs I believe)

apps.tv creates this on the fly based on the TiVo's ID which is sent in the headers.

wmcbrine
01-21-2008, 04:21 AM
Thanks. Interesting... I may have overlooked a port 80 connection attempt.