You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tomcat.apache.org by bu...@apache.org on 2002/04/26 21:16:27 UTC
DO NOT REPLY [Bug 8566] New: -
Filter using ServletRequestWrapper breaks RequestDispatcher.forward
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=8566>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND
INSERTED IN THE BUG DATABASE.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=8566
Filter using ServletRequestWrapper breaks RequestDispatcher.forward
Summary: Filter using ServletRequestWrapper breaks
RequestDispatcher.forward
Product: Tomcat 4
Version: 4.0.3 Final
Platform: PC
URL: http://qlink.queensu.ca/~9cw4/loop-bug.zip
OS/Version: Windows 9x
Status: NEW
Severity: Major
Priority: Other
Component: Catalina
AssignedTo: tomcat-dev@jakarta.apache.org
ReportedBy: cwolfe@globetrotter.qc.ca
http://qlink.queensu.ca/~9cw4/loop-bug.zip contains a simple web-app that
produces the bug and implements a simple workaround.
* Problem:
When a filter uses a ServletRequestWrapper to modify the request in doFilter,
any changes made by ApplicationDispatcher will be completely ignored.
* Workaround:
When getRequest is called on the wrapper, its getServletPath reverts to calling
super.getServletPath(). This works correctly in 4.0.3, but could be broken
easily in future versions.
* Context:
In org.apache.catalina.core.ApplicationDispatcher:
doForward(request, response) calls wrapRequest for HTTP path-based forwards.
wrapRequest walks up the request chain past any ServletRequestWrappers and
inserts a new ApplicationHttpRequest before the first wrapper, this new request
is returned.
doForward modifies the returned request to match the target of the forward.
* Discussion:
Why does the insertion need to take place before the first ServletRequestWrapper?
Considering the JSP spec specifically talks about using ServletRequestWrapper in
a Filter to redirect requests, it breaking forward is a serious problem.
Removing the chain walking code would seem to be the obvious fix, but I am
worried about breaking includes or other behavior.
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>