Re: i28 proposed replacement text

On tis, 2008-06-03 at 19:23 +0200, Julian Reschke wrote:
> Well, unless I'm missing something, it will be hard to send from a 
> servlet (hey, Servlet EG, are you listening...?).

Content-MD5 either needs the server to buffer the response, or sent in
the trailer. Same for Content-Length except that it can't be sent in a
trailer.

Claiming that those isn't usable for a servlet is plain wrong. It's all
an implementation detail, and how feasible it is depends primarily on
the size of the response.

> Well, what would they contain in case of a truncated response? Surely 
> not the length/digest of the actual response, because that wouldn't help 
> the client finding out about the truncation...

Indeed.

> Maybe something like "final-status" as a new response header would make 
> sense. That way, a server could send an initial 2xx, start sending 
> content, and in case of internal errors could at least signal that 
> something went fatally wrong...

Problem there is that recipients are not required to care about trailers
and those who don't will misread the response as 100% successful.. so
you are better off simply closing the connection in the middle of the
response and log the error locally. The next-hop will notice the error,
but there is no guarantee the final recipient will..

The scope of this WG is to clarify HTTP/1.1 and correct errors, not
HTTP/1.2 (or 2.0) fixing the shortcomings of HTTP/1.1.

Regards
Henrik

Received on Tuesday, 3 June 2008 18:08:09 UTC