- From: Scott Lawrence <lawrence@agranat.com>
- Date: Wed, 30 Apr 1997 15:54:22 -0400
- To: Robert Herriot <Robert.Herriot@eng.sun.com>
- Cc: ipp@pwg.org, http-wg@cuckoo.hpl.hp.com
SDL> From lawrence@agranat.com Wed Apr 30 12:05:41 1997: SDL> Correct; the 'Transfer-Encoding: chunked' applies to all of the HTTP SDL> message body. For completeness, my companys' server implementation SDL> does support chunked encoding of the entire multipart/* body part, SDL> but we think it doesn't make much sense (because it is redundant) so SDL> that support may be compiled out to save code. >>>>> "BH" == Robert Herriot <Robert.Herriot@Eng.Sun.COM> writes: BH> Why is chunking redundant? It seems important for the case where the BH> client doesn't know the length of the data at the beginning of the BH> transmission. Without the Transfer-Encoding of chunked, HTTP/1.1 seems BH> to require a Content-Length for a client transmission (as a client, it BH> cannot close the connection and multipart/byterange doesn't seem BH> appropriate for clients to send). Not in general! Generating a multipart/* body (which has mechanisms already in it for describing the length of each part), and then wrapping all that in an HTTP 'Transfer-Encoding: chunked'. It's not that we think it won't work or even that it would never be the right thing, just that in the application environments we are targetting it is too expensive in code complexity to be supported by default (code size is _very_ important to our customers; our 1.1 server can be under 20K). -- Scott Lawrence EmWeb Embedded Server <lawrence@agranat.com> Agranat Systems, Inc. Engineering http://www.agranat.com/
Received on Wednesday, 30 April 1997 12:57:24 UTC