> -----Original Message-----
> From: Henrik Nordstrom [mailto:hno@hem.passagen.se]
> Sent: Thursday, 8 February 2001 11:14 AM
> To: Robert Collins
> Cc: squid-dev@squid-cache.org
> Subject: Re: rfc 2616 query
>
>
> From my understanding yes, but it is of quite limited value as in most
> cases the full length is known.
Good. That was my understanding too. (Range requests must specify the
full entity length in the content-range header, but the content-length
!= entity length, so we don't mange the range reply when we remove the
content-length).
> When you know the length, plain encoding is more efficient for all
> involved (unless you also want to compress the data).
Two cases: One is compression, the other is with operations on the data
which may (deterministicly) change the body - ie Joe & Moez's
acceleration url rewriting. While the second case is rare, I don't want
to confuse browsers because of it :]
Also other things allowed but not recommended, such as content-type
transformations will need to interact cleanly with transfer-encoding.
Rob
Received on Wed Feb 07 2001 - 18:30:44 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:13:28 MST