You're right, and that's exactly what I do 90% of the time. I was trying to see if I could provide a 100% client-side, no-proxy solution by way of JSON callbacks for a Silverlight library, and the proxy strategy works great, I was just pushing there. And it works for any call that doesn't need auth, at least. But it seems less than useful in the grand scheme of things.
On Jan 13, 3:33 pm, "Steve Brunton" <[email protected]> wrote: > On Tue, Jan 13, 2009 at 2:16 PM, Dimebrain <[email protected]> wrote: > > > I was hoping to be able > > to keep everything on the client-side; can you use a server proxy > > without leaving the client, i.e., so that the json callback has > > somewhere to call back to? > > If you are using a proxy back through the originating server why would > you need a callback? At that point in time you can just use straight > up Ajax calls to get JSON or XML back to the browser and all all the > web 2.0 magic voodoo that you want to. This is exactly what I did for > some search stuff I was messing around for ${work}. Make > /proxy/twitter/search a proxy to search.twitter.com and then make all > Ajax calls to my server which then proxies to the search.twitter.com > server and passes back the JSON results I'm looking for. Granted, > nothing I'm doing for search needs authentication, but a proper setup > should return all the headers through for a 401 to hit the browser. Of > course at the point in time you once again open yourself up to the > fact that your server is going to get the username/password before it > gets passed through to Twitter and [insert evil things you can do with > that here]. > > -steve
