Re: mid-course errors

"Henry Sanders (Exchange)" <henrysa@EXCHANGE.MICROSOFT.com> pointed
this out to me:
    > In the case of a server sending a chunked response intending to
    > put the content-length in the footer (which is allowed by
    > section 3.6 of RFC2068 (Transfer Codings)), it seems reasonable

    Umm, where do you see this in section 3.6? I read it as saying that
    you're only allowed to send entity header fields in the footer
    which are explicitly defined as appropriate, as I don't see any
    wording anywhere explicitly defining Content-Length: as
    appropriate. I know our server doesn't check for Content-Length: in
    the footer, and I don't think our client will either.

I may have misread this section.  

But, it says:

   The chunked encoding is ended by a zero-sized chunk followed by the
   footer, which is terminated by an empty line. The purpose of the
   footer is to provide an efficient way to supply information about an
   entity that is generated dynamically; applications MUST NOT send
   header fields in the footer which are not explicitly defined as being
   appropriate for the footer, such as Content-MD5 or future extensions
   to HTTP for digital signatures or other facilities.

Content-length is definitely an entity-header (see section 7.1), and I
think it is clearly "information about an entity that is generated
dynamically" (although whether "that is generated dynamically" refers
to the "entity" or the "information" is ambiguous).

What's interesting is that there seem to be NO fields in RFC2068
that are "explicitly defined as being appropriate for the footer",
including Content-MD5.  I searched the document for the word "footer",
and (besides section 3.6) it appears in only one other section (8.2).
There is no mention at all in the section on Content-MD5 (section 14.16).

Does your server or client actually have a list of headers that it
*does* allow in the footer?  Where did this list come from?

I think we have some document-editing to do.  Either we have to
remove the "MUST NOT" from section 3.6, or remove the example of
Content-MD5, or add a list of "explicitly allowed in the footer"
headers.

Anyway, it seems quite reasonable to allow Content-Length to appear
in a footer.  And it seems to solve the "midcourse correction"
problem, in most cases.

-Jeff

Received on Tuesday, 15 April 1997 13:55:55 UTC