> > My only other observation is that any filters added in the subreq > > must only be added to the top of the rnew->output_filters > > chain or must be cleanly removed before the subreq returns > > otherwise the main chain will get corrupted. I haven't seen > > this happen so far, so maybe this is just a hypothetical > > concern. > > It isn't hypothetical, but it also isn't a bug we want to protect > against. The thing is that the filters added in the subrequest by > definition must be resource filters, which means that they must be added > above protocol and connection filters. If you are adding either a > protocol or connection filter in a sub request, then you are doing > something incredibly wrong, and we want the server to fail.
Absolutely. The case I was actually thinking about but didn't clearly explain was something like say mod_deflate, which if it were "AP_FTYPE_CONTENT_SET" would be above the proto chain passed by next_filter. Even so I don't know of any subreq filter that would get placed after this. Allan