On Jan 29, 2012, at 2:53 PM, Stefan Fritsch wrote: > On Wednesday 25 January 2012, Jim Jagielski wrote: >> Looking over the code, impl as a hook seems more "isolated", >> rather than the current impl which is intrusive (which is >> part of what we're trying to avoid, aren't we?) > > OK, patch is attached. This needs review/testing for Windows. > <insert_network_bucket.diff>
Just a quick scan but: +AP_IMPLEMENT_HOOK_RUN_FIRST(apr_status_t, insert_network_bucket, + (conn_rec *c, apr_bucket_brigade *bb, + apr_socket_t *socket), + (c, bb, socket), DECLINED) + ... + ap_hook_insert_network_bucket(core_insert_network_bucket, NULL, NULL, + APR_HOOK_REALLY_LAST); looks wonky...