Ticket #507 (defect)

using InternalRedirect('foo', params=params) fails, because unlike HTTPRedirect, it does not urljoin the path argument with the request's current path.

Here is a short diff adding urljoin:

Index: _cperror.py
===================================================================
RCS file: /cvs/third-party/CherryPy/cherrypy/_cperror.py,v
retrieving revision 1.1.1.1
diff -u -p -b -r1.1.1.1 _cperror.py
--- _cperror.py 27 Mar 2006 05:31:03 -0000      1.1.1.1
+++ _cperror.py 12 Apr 2006 16:23:38 -0000
@@ -31,11 +31,16 @@ class InternalRedirect(Exception):
     def __init__(self, path, params=None):
         import cherrypy
         import cgi
+        import urlparse
         request = cherrypy.request
         
         # Set a 'path' member attribute so that code which traps this
         # error can have access to it.
-        self.path = path
+        #
+        # Note that urljoin will "do the right thing" whether url is:
+        #  1. a URL relative to root (e.g. "/dummy")
+        #  2. a URL relative to the current path
+        self.path = urlparse.urljoin(request.object_path, path)
         
         if params is not None:
             if isinstance(params, basestring):