To follow up to Paul's message, the situation of an empty and/or missing body is not well-specified in SIP, and a SIP parser should be careful to "be liberal in what you accept". In particular, if the observed or specified content length is zero, it *might* be that there is no body, or it *might* be that a body is present (of the type specified by Content-Type), but its encoding has zero length (the significance of which depends on the Content-Type). If the processing of the message would succeed under one or the other of these conditions, it should succeed.
In addition, if the processing of the message does not require knowing the contents of the body, it should succeed even if the body is unexpected, has an unknown Content-Type, or is invalid for its Content-Type. (Although it seems reasonable to be stricter regarding multipart types, and perhaps require that the multipart structure per se is valid.) I am reminded of the sipX system, which when genereating error responses, would add a body of type message/sip, which was the request that caused the error. This allowed the UAC to see in the error response what the request looked like at the point where the error was generated. We never found any UAC that failed because of the unexpected body (since no defined error responses at the time had specified bodies). Dale _______________________________________________ Sip-implementors mailing list Sip-implementors@lists.cs.columbia.edu https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
