I'm not sure how to read the RFC but instead decided to have libcurl be 
the reference. It does send a trailing crlf, and now so does 
Webclient-Core-ar.83.

Cheers,
   - Andreas

On 1/17/2011 9:08 PM, Chris Muller wrote:
> Section 2.10 of RFC 2045 defines "Lines" as:
>
>      '"Lines" are defined as sequences of octets separated by a CRLF sequences.'
>
> When I try to upload a file to a KomHttpServer via a WebClient, I get
> a timeout because the last line of
> HttpRequest>>#multipartFormFieldsDo: expects a CRLF after the last
> body part and boundary indicator of a multipart message.
>
> However, WebClient's WebUtils class>>#encodeMultipartForm:boundary:
> does not emit a CRLF after the final boundary, causing the
> KomHttpServer to timeout waiting for it.
>
> I must admit, RFC 2045's definition of a "line" is ambiguous, to me,
> in terms of whether the boundary indicator after the LAST body part
> should be "separated" by a CRLF sequence.
>
> Any HTTP experts out there know the answer to that?  I need to know
> whether WebClient or KomHttpServer should be fixed..
>
> Thanks,
>    Chris
>
>