> > Well, after looking at the algorithm (and trying with a telnet), I think
> the
> > flag does the job. However, this is likely to be a capability which will
> be
> > going away when I rewrite the connector (since it adds complexity
without
> > adding anything useful - see below).
>
> After the 'allowChunking=false' flag, I still got chunked responses. I can
> make a simple test case if you would like that.
Sure.
I tried a normal request to a servlet which didn't set the content length,
and it worked as expected.
> > Chunking is not optional for HTTP/1.1 compliance. If you support
HTTP/1.1,
> > you have to support chunked encoding (it's quite easy to do, actually,
as
> > it's a very simple encoding).
> >
> > Remy
>
> The company I work for controls both the client and server. It is a
> worldwide network for message delivery.
> We have 3 web server constraints.
> - Support for HTTP 1.1.
> - Support for Keep-Alive.
> - We do not want chunked encoding. Streamed responses will have the
content
> length in the header.
If you have specific need about features, the best would be to modify an
existing connector.
Remy
