On 17/3/03 20:15, "Stefano Mazzocchi" <[EMAIL PROTECTED]> wrote:
>> On another thought... The cache should store unicode characters "as is", not
>> bytes, as those might change for the same request URL depending on the
>> different headers in the request...
>
> Uh, another good point.
Ok, in the light of what Dirk said regarding to the "Accept-Encoding", I'd
propose an interim solution that is not fully "http" compliant, but at least
is better than what we've got at the moment...
I believe that we should go with what you said in another reply to this
thread:
> What about adding a 'content encoding' attribute to the 'pipeline' instead?
Right now, to make things work in 99.9% of the cases, I'd say that we add
this to the pipeline, and right now, we enforce this onto the client. So, no
matter what he asked in the different "Accept*" headers, we deliver them
what _we_ want...
As negotiating the encoding/type/language is something that will also affect
widely our cache, I propose to defer this to Cocoon 2.2 (or later) when the
problem can be analyzed thoroughly in more details...
Now, does anyone have suggestions on how to retrieve the "charset" value out
of <map:pipeline> ??? Where would be a nice place in our API?
On a side not, currently, "AbstractTextSerializer" relies on Xalan as its
way to generate content, but, IMO, there could be other text serializers not
needing it (ok, it's JAXP! :-)
Pier
