On 11/3/10 7:06 AM, Jonas Sicking wrote:
No, my concern is that browsers will implement this, and then sites that
haven't updated their jquery, and probably never plan to do it, will start
using the new stuff browsers have implemented.
But won't that always fail? If the author either sets .responseType
when jquery doesn't expect it, or the author uses .response when
jquery hasn't set .responseType?
In the particular case of jquery, yes, unless there are some try/catch
blocks around.
With other things, it might not be so simple. Looking at
http://prototypejs.org/assets/2009/8/31/prototype.js it seems they
try/catch around their .responseText stuff and turn those exceptions it
into events....
Yes, it means that pages that use old code can't use the new features
together with that code, but I don't see that as a big problem in this
case.
My problem is that it seems, to me, to be really easy to get into
situations where things _seem_ to work, but don't really, or don't with
a slight change in totally unrelated code. Since jquery always gets
responseText that's not the case there in that you'd get an exception at
that point. But prototype is catching such exceptions, and again my
concern is cases where different codepaths end up assuming different
things from the same XHR in hard-to-debug ways.
I might be wrong, of course. But I think we'd be creating a debugging
nightmare for web page authors.
-Boris
