https://issues.apache.org/bugzilla/show_bug.cgi?id=56035
--- Comment #12 from Yann Ylavic <[email protected]> --- (In reply to maapu from comment #11) > @Yann > But why do we need to call ap_die incase of ap_is_HTTP_VALID_RESPONSE > returns false. Why cant we just call ap_finalize_request_protocol? Will it > miss some cases? ap_die() ensures the client will receive an error response in any case, should an handler return a non-HTTP status without responding by itself (or calling a filter), "Internal Server Error" will be used. For the logging bug however, I think the proposed patch(es) don't do the right thing, ap_process_request() should always set r->status to HTTP_OK before calling ap_die(), or the latter may think it's a recursive error (while it is never here). So maybe the correct way to handle this is to check whether the status is still HTTP_OK after ap_die(), in which case a response has already been sent to the client before, and the original r->status is relevant (and to be restored). This new patch follows. Regarding the double response body, as Eric said, this comes from (mod_)php just ignoring the error, no httpd's level patch can resolve this IMHO. -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
