https://bz.apache.org/bugzilla/show_bug.cgi?id=58551
--- Comment #11 from Mark Thomas <ma...@apache.org> --- I remain unconvinced that this feature is a) necessary or b) a good idea. (In reply to romain.manni-bucau from comment #10) > Why Tomcat couldn't get this? The code itself is very limited. Implementing this feature is not as simple as porting the Cassandra patch you referenced. It would require changes all the way through the I/O layers from SocketWrapper all the way up to the Request since whether or not a request is secure and the protocol in use would become a property of the socket rather than the connector. Implementing this feature would also break the current orthogonality between the secure, protocol and SSLEnabled properties of a connector. The more I think about it, the more this aspect of this proposal concerns me. The current implementation of these features is well understood and offers a great deal of flexibility for a wide range of applications. Adding automatic plaintext/TLS switching would muddy the waters considerably. It is also worth noting that there is a standard for this: RFC2817 and that that standard is largely unimplemented by web servers and clients due to numerous issues such as those articulated here: https://bugzilla.mozilla.org/show_bug.cgi?id=276813#c7 I don't see why adding an additional HTTP connector (given that an HTTPS is going to have to exist to handle the TLS traffoc) is so hard. -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
