https://issues.apache.org/bugzilla/show_bug.cgi?id=57215
--- Comment #1 from Mark Thomas <ma...@apache.org> --- This is going to get messy. The Javadoc for HttpServletRequest.getContextPath() says the container should not decode the returned value. Where this gets 'interesting' is when the URI is not normalized and is encoded. For example, what gets returned for a request to "%2Ffoo%2F%2E%2E%2Fpath"? Is it: "%2Fpath" ? "%2Ffoo%2F%2E%2E%2Fpath" ? Something else? We know (from the mapper) how many '/' characters to include in the context path. The current approach of simply searching that many '/' characters down the request URI ignores issues of normalization and encoding. Doing that counting in a normalization and encoding aware manner is probably the answer but that is non-trivial to say the least. Fixing this bug might not solve the problem you are seeing - particularly since the unit tests you provided are using the incorrect value for the expected context path. You should probably be using ServletContext.getContextPath(). -- 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
