As you might know, there is much discussions around modifications that shipped with Safari in iOS 6, including a 1 connection per domain bug or feature.
I’m not sure it’s even real because I can not test it myself and it seems a bit too dangerous as a modification.
the iPhone on WPT Dulles is still on iOS 5.1 and as long as the iOS6 is not widely adopted, it’s all good, but would it be possible to have a 2nd iPhone with iOS 6, or even an emulator, to check if the network is used differently ?
I’ll see if I can get something set up but it may not really answer the question. WPT does testing with a UIWebView embedded in a test app so it will be able to show you how an embedded web view performs (like in the twitter client, etc) but not how the native Safari browser performs (if they behave differently).
I expect we’re largely going to have to wait for more information as people test it more thoroughly manually to know for sure. Suffice it to say, it looks like they shipped at least of couple of critical bugs (caching POST requests - javascript - Is Safari on iOS 6 caching $.ajax results? - Stack Overflow and the connection behavior being at the top of the known list right now).
Luckily iOS 6 also included support for remote dev tools so it should be easy enough to look into both cases.
I saw those links, and for now everyone seems to look for the exact behavior.
I did not saw the “Connection per hostname” = 0 in browserscope. Weird, I guess the way they test it is not adapted to some Safari optimization ? I could not find the explanation on how they are doing the test.
I forgot the monitoring on WPT was made from a WebView, but it’s worth testing, if anyone has time