Simran, Just trying:
GH> however, any changes made to $r->document_root GH> would persist for the entire life of the child, not just the life of GH> the current request. for that reason, you were expected to always GH> restore the value of $r->document_root to its original state. GH> GH> see recipe 4.3 in the cookbook > Thanks for those... those are the exact ones i have followed. Can it be that a subrequest (issued (by chance?) to the same child) is messing things up? You might try adding a warn to the CleanupHandler. Just my 1.5c --Frank
