Re: interesting note

On Fri, 9 Feb 2001, Robert Collins wrote:

> reading /. they noted this page
> http://www.w3.org/TR/2001/NOTE-cuap-20010206. Item 1.7 there intrigues
> me - is the implication that caches should buffer the entire entity
> before beginning transmission so that the status code is known? Or is
> aborting the transfer (for example: on a chunked response drop the
> connection without terminating the chunked data in any way acceptable)?

W3C document misquotes RFC 2616. The paragraph they point to refers
only to the situation when a cache is trying to serve a hit to the
client using a previously cached partial response. W3C text seems to
imply that RFC requirement is "generic" (i.e., applied to all
conditions). It is not.

Note that in cases where Content-Length is available, the browser can
detect aborted transfer.

Alex.
Received on Thu Feb 08 2001 - 16:13:34 MST