W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2014

Re: #445: Transfer-codings

From: Ilari Liusvaara <ilari.liusvaara@elisanet.fi>
Date: Fri, 4 Apr 2014 21:33:16 +0300
To: Matthew Kerwin <matthew@kerwin.net.au>
Cc: "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>, Mark Nottingham <mnot@mnot.net>
Message-ID: <20140404183316.GA24485@LK-Perkele-VII>
On Sat, Apr 05, 2014 at 03:52:55AM +1000, Matthew Kerwin wrote:
> On Apr 4, 2014 9:06 PM, "Ilari Liusvaara" <ilari.liusvaara@elisanet.fi>
> wrote:
> >
> > Are the compression contexts chained? If so, using multiple compressions
> > one could use some serious memory (especially as window sizes are not
> > limited).
> 
> No, each DATA is encoded independently. I guess that wasn't clear. Isn't
> that how T-E worked in 1.1?

No. In HTTP/1.1, gzip (just as example) T-E works in either of two ways:

- Gzip the whole response and then chunk the gzipped stream.
- Gzip the whole response, send result as-is and close the connection on
  end.

Also, HTTP/1.1 T-E supports multiple transforms (sometimes useful for e.g.
applying transforms that enhance compressiblity):

Transfer-Encoding: foo, gzip, chunked

(If foo is some filter that tries to make data more compressible by gzip).
The dataflow goes (raw data) -> foo -> gzip -> chunk -> (output)


-Ilari
Received on Friday, 4 April 2014 18:33:41 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:29 UTC