[DBMandrake]

Ha that testbuild is missing the srcvers by accident if you ask me. DBMandrake can you verify that this change really speeds up video target detection?

Finally got a chance to try this. Can't detect any difference I'm afraid, either for better or worse.

Now granted I didn't try the actual oct 5th build - what I did was remove the srcvers variable from the announcement on my existing 13.2 custom build so that the only thing that has changed from my current setup is that one variable. I think you'll agree that any difference in detection must be in the announcement as both Gotham and Helix have the same 10 second refresh, correct ?

What I did notice is that whether it is detected as video immediately or not is highly erratic in the first place. I have both a Mac running my 13.2 build, and a Raspberry Pi on the network running the latest Raspbmc which has the 0xF7 fix in it - both broadcasting the exact same announcements, and before I even made any changes to srcvers I noticed that most of the time the Mac was immediately being detected as video while the Pi was detected as audio for up to 10 seconds.

This was on two iOS 8 devices even after they were rebooted with wifi turned off. Then for no good reason the Mac then started being detected as audio for up to 10 seconds. After removing srcvers I saw instances of both instant video detection and slow video detection...

So I have to wonder whether AussieFries switched back and forth between the two test builds multiple times and tested a sufficient number of times on each (including rebooting iOS devices) to draw a firm conclusion, as my results for speed of detection as video were very erratic with or without srcvers...

The Pi was never detected as video immediately but the Mac often was even with identical announcements. Perhaps a small difference in the refresh behaviour between Apple's Bonjour and Avahi ?