- From: Jonas Sicking <jonas@sicking.cc>
- Date: Wed, 05 Apr 2006 03:57:37 -0700
- To: Stéphane Sire <sire@intuilab.com>
- Cc: public-webapi@w3.org
Stéphane Sire wrote: > > Hi, > > I have been quickly through XHR draft (version dated 20th March 2006 on > CVS) and I have the following two remarks: > > - nothing is said about how to create/manage timeout; I guess the idea > is to use the external timer API (for instance in Window object) and > then use cancel to end a timed out request... Apparently HTTP doesn't define timeouts so I think it'd be bad for us to add any requirements. So yeah, I think creating your own timers or relying other mechanisms to abort requests is the way to go. I believe that in mozilla we will enable the stop button when the request is made and pressing that will abort the load. > - what happens if the browsing context is destroyed, is the request > canceled or not ? is the onreadystatechange function called ? I'm not sure we want to dictate to browsers what to do here. I would not be surprised if some browsers disabled all scripts once you leave a page. / Jonas
Received on Wednesday, 5 April 2006 10:57:50 UTC