Roland, Thats a good argument. I would like to add that even if the server is RFC 2965 compliant or understands multiple cookie specs in the same header (which I highly doubt), there is a high probability that it is backward compliant with old specification and will understand old style formatted cookies. In addition to this, we also send a Cookie2 header which informs the RFC 2965 compliant server that the client understands Set-cookie2 cookies.
Samit On 8/29/05, Roland Weber <[EMAIL PROTECTED]> wrote: > Hi Samit, > > > There is some ambiguity regarding formatting cookies which include > > both new-style and old-style cookies. I don't think it is appropriate > > to format cookies differently in the same header. Therefore, if there > > is an old-style cookie present amongst the cookies to be formatted, I > > delegate all formatting to RFC2109 spec. But I may be wrong on this, > > let me know ur suggestions. > > From RFC 2616, one could infer that multiple Cookie: headers must be > concatenable into a single one. But there are two arguments against > this conclusion: > > 1. Non-conforming Netscape style cookies (comma in date). > 2. The "single-cookie-header" option has been introduced to HttpClient > for interoperability with servers that are *not* fully conformant > to the various HTTP related specs. > > If the server isn't able to understand multiple cookie headers, we > should not expect it to understand multiple cookie specs in a single > header either. I think your approach is appropriate. > > cheers, > Roland > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
