https://issues.apache.org/bugzilla/show_bug.cgi?id=39740
--- Comment #8 from William A. Rowe Jr. <wr...@apache.org> 2011-02-08 19:12:35 EST --- First, those comments are nonsense since 1999, as the URI is a quoted construct not subject to sgml parsing, and any sgml parser corruping quoted sgml tags is broken, end of discussion. Your post is similarly nonsense since ';'s are allowed to be passed, the only question is what is made of them. Second, http://tools.ietf.org/html/rfc3986 is the only relevant references which http considers at the protocol layer. "HTML 4.01 Specification" was the origin of the '&' construct - http://www.w3.org/TR/html401/. It's reasonable by 3986 to put anything in a query arg as long as it is not a literal '#' symbol. The concept of query key=val pairs originates in CGI, which was formalized by W3C, http://www.w3.org/TR/html401/interact/forms.html#h-17.13.4 without deprecating the old '&' recommendation, so nobody adopts this change. But it is entirely up to the application, so whichever decision the servlet spec authors adopt is acceptable. If they conform to what the user agents have offered ('&' only) that's fine too. What the W3C offered here is certainly not worthy of being called a 'specification', as it was not specific. To Christopher, you are getting your path and query segments confused. In the path part, /foo/bar is not /foo/bar;x=1 and that's made very clear in the spec. And in your example /foo/bar;uid=1 is not /foo/bar;uid=2, something that is very deliberate in presenting per-session data. -- Configure bugmail: https://issues.apache.org/bugzilla/userprefs.cgi?tab=email ------- 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
