On 10/27/10 3:40 AM, Boris Zbarsky wrote:
On 10/27/10 3:24 AM, Darin Fisher wrote:
So, it sounds like we just need to present you with some concrete
examples of XHRs being used to fetch large responses as Text or XML, and
then you will be convinced?
Uh... no. I'm sure these happen. The question is whether their existence
is sufficient grounds to impose additional burdens on other uses of XHR.
And specifically for the case you cited (large XML file that you process
via XML) Webkit is already storing two copies of the data: the
responseText and the parsed DOM. I'd be very surprised if the latter is
much smaller in terms of memory used, and somewhat surprised if it's
smaller at all.
So for this specific use case, in fact, if you just stored the raw bytes
and the parsed DOM you would be no worse off than you are now (and this
is in fact what Gecko does in this situation).
The double storage issue really only comes up for cases when
responseText is being used, not responseXML, as far as I can tell.
-Boris
