On Wed, Mar 06, 2002 at 11:30:49AM -0800, Ryan Bloom wrote:
> Index: server/core.c
> ===================================================================
> RCS file: /home/cvs/httpd-2.0/server/core.c,v
> retrieving revision 1.159
> diff -u -r1.159 core.c
> --- server/core.c 6 Mar 2002 18:03:19 -0000 1.159
> +++ server/core.c 6 Mar 2002 19:52:52 -0000
> @@ -4040,7 +4040,7 @@
> AP_FTYPE_NETWORK);
> ap_subreq_core_filter_handle =
> ap_register_output_filter("SUBREQ_CORE",
> ap_sub_req_output_filter,
> - AP_FTYPE_HTTP_HEADER);
> + AP_FTYPE_CONTENT + 5);
> ap_old_write_func = ap_register_output_filter("OLD_WRITE",
> ap_old_write_filter,
> AP_FTYPE_CONTENT - 10);
Should we add a new filter type (say AP_FTYPE_CONTENT_SET) where
SUBREQ_CORE, mod_deflate, mod_headers, etc should go? It would
be in between AP_FTYPE_CONTENT and AP_FTYPE_HTTP_HEADER (which
should be renamed to AP_FTYPE_PROTOCOL). The idea is that the
content-type should be finalized at this filter level. The name
sucks though. -- justin
