Alex Hill

Testing in IE shows the same native loading behaviour as Chrome and Firefox: the address bar isn't updated until the response is received. To sum up my position, I feel...

Hmm, yes, it does. I think it changes when DNS resolution fails, so pretty much instantly. Try this [link with 5s delay added](http://deelay.me/5000/http://www.bubblews.com/assets/images/news/418949094_1382795658.jpg) to see what I mean. Safari changes...

Sorry it's taken a while to follow this up. > Do you think it would work if we explicitly set the referrer header with the old URL for fallback requests?...

No worries. I've opened a PR just to add an XHR abort to the top of the popstate handler here: #395

And here's a PR fixing the HTTP referer bug without moving pushState: #397

Hi @mislav, I'd like to restart this discussion and hopefully get this change in. I've just rebased against current master and cleaned up a bit. Since you merged #395 this...

Hi @olemoign, can you confirm you're definitely running my code – no caching issues etc? Have a look for yourself: [the only place cachePush is called is in the AJAX...

Right you are, thanks for the clarification. Didn't read closely enough, sorry :) I've just tried moving some things around in the popstate handler, so that `cachePop()` is only called...

Yeah, this is pretty painful. Draining should respect the update_config, or have an equivalent option. The `pause`/`update --force` method is pretty good, but it results in a lot of churn...

@thaJeztah just a general question - are issues like this generally amenable to being addressed by PR from a member of the community? A fix would raise questions of backwards...