DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT <http://issues.apache.org/bugzilla/show_bug.cgi?id=28839>. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE.
http://issues.apache.org/bugzilla/show_bug.cgi?id=28839 The Coyote connector publishes empty URIs to catalina ------- Additional Comments From [EMAIL PROTECTED] 2004-05-11 00:35 ------- The problem (as I see it) is that the request I'm making (see the attached access log: id=11463) should NOT be routed to the default servlet (as it is). I expect (and tomcat does for the deprecated HTTP connector) that the request be routed to the library servlet (initialization of which you can see in the Host log: id=11462). For unknown reasons, coyote seems to publish the URI as empty (info gleamed from the host file) while I expect '/library/services/library' (what I actually hit, and whose information is visible in the access log). I extrapolate this last part per the details listed in the aforementioned Host log file. I'm initially hoped that this was simply a configuration issue, but the same server.xml (attachment id=11460) file works when the HttpConnector is specified in place of the CoyoteConnector and does NOT work when the CoyoteConnector is used. Hence, if it is a configuration issue, then this bug becomes one of needing better documentation as to what the configuration file(s) need to contain in order for the CoyoteConnector to operate properly. So the problem, initially, was that, after upgrading to tomcat-4.1.30, when I hit a web page generated by the library servlet, I was redirected to the "/" URI. This took place no matter what URI I was hitting. Backtracking to tomcat-4.1.24 via setting CATALINA_HOME to the older base, leaving CATALINA_BASE as it was, resulted in a perfectly operating system. Leaving tomcat-4.1.30 configured but using the deprecated HttpConnector resulted in a well behaved web application (as I have already mentioned). So, changing one line in server.xml resulted in the system either working or not working depending on whether that line indicated the http or coyote connector. Workaround: use the older HttpConnector. So, my question is, why? Why does using the coyote connector cause the system to not route requests to the appropriate web application/servlet? Why do all requests route to the default servlet? The only answer I was able to come up with (via investigating the log files) is the clue that the standard host class is routing requests based on an empty URI, when I knew that those URIs were not empty strings. As URI interpretation is done by the coyote connector, my best guess is that the coyote connector is broken WRT the URIs it publishes to the engine. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
