Anne van Kesteren wrote: > On Fri, May 16, 2014 at 12:20 PM, Tim Taubert <[email protected]> wrote: >> Calling the whole idea of <a ping> "disturbing" makes it sound like we >> would introduce a whole new concept, we just provide a saner way to do >> things that lots of web pages want. There is no obvious disadvantage to >> the user from my POV here. > > It should be better as closing a document is no longer halted by > synchronous XMLHttpRequest. Are we implementing them in such a way > they can be transmitted even if the window goes away? So performance > for the user is optimal and developers don't feel forced to slow > things down?
Yes, <a ping> is implemented only to ping when the user clicks a link and navigates away from the current document. Thus transmitting the ping when the inner window goes away is a requirement. The outer window going away isn't a problem either because we create independent http channels/requests that will be held alive by the networking subsystem until they succeed or time out. - Tim -- Tim Taubert Engineering Manager, Firefox @ttaubert _______________________________________________ dev-platform mailing list [email protected] https://lists.mozilla.org/listinfo/dev-platform
