websockets branch needs testing

[John Wiseman]

I've pushed a websockets branch that uses a persistent websocket connection to get the latest mavlink messages, which should be lower overhead than doing full HTTP GETs.

I've tested it in desktop Safari and Chrome, and iPad Mobile Safari and Chrome, but my iPhone is jailbroken and stuck on iOS 5.1.1, and that version of Mobile Safari implements an incompatible version of the websocket spec.

If anyone can test it in Mobile Safari for iOS 6.0.1 or any other mobile browsers I'd appreciate it.  I'm curious whether it works and whether the position and orientation updates seem any smoother.  You'll need to "pip install ws4py" as well.

If we find incompatible browsers I can add some runtime logic to fall back to GETs.

FYI that branch also has two other changes: 1. I removed flask and werkzeug and ported everything to cherrypy because it got too complicated to glue everything together with WSGI  2. I changed the API URLs--they're all under /mavelousapi now and I renamed a couple (like /mavlink -> /mavelousapi/latest_messages).

And at the moment offline mode does not work in this branch.

Thanks!

John

[Richard Miller]

Is the Websockets branch still of interest?  I was thinking of trying to get a PFD app running (possibly on iOS) and if everything were standardized on Websockets the HTML/JS/CSS devices could run with the same data.  PFD on my iPhone, Map on an iPad, maybe an HSI on another device, with redundancy all around.

I wrote an iOS websockets <-> Python app last year, but have been away from it a while, so not sure whether websockets is still the way to go.

Richard. 

[John Wiseman]

It is still of interest, though I'd like to measure the actual benefit it gives us.  I also found websockets to be somewhat of a moving target with respect to standards: If I use a version of websockets that works with iOS 6.x it doesn't seem to work on iOS 5.x.  I got as far as wondering if XHR streaming (is that Comet?) would be more compatible but haven't had a chance to go deeper.

John

--
You received this message because you are subscribed to the Google Groups "mavelous" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mavelous+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.