I've written a patch that seems to address the problem.
I just wanted others to have the option to apply it. I think it's
ridiculous to have to break apache, an RFC compliant server to be able
to view a website. But what am I going to do about it? :) The truth is
the patch should remove ANY entity-header variable. But I only chose to
deal with the Content-Length variable in the short term.
Comments are welcome as always.
Enjoy,
Geff
On Mon, 2002-08-12 at 11:47, Graham Leggett wrote:
> Geff wrote:
>
> > I'd also like to say that this issue doesn't just affect the 'no-cache'
> > case but it also seems to affect the case when the cache decides it
> > needs to revalidate a resource with the original server. The 'no-cache'
> > method is just a convenient way to demonstrate the problem and force the
> > cache to try to revalidate the resource.
>
> From what I understand this is caused by some buggy versions of IIS v5
> which send Content-Length: 0 along with 304 responses.
>
> Regards,
> Graham
> --
> -----------------------------------------
> [EMAIL PROTECTED]
> "There's a moon
> over Bourbon Street
> tonight..."
>
*** /home/boing/apache_1.3.26/src/modules/proxy/proxy_cache.c Mon Aug 12 17:20:58 2002
--- proxy_cache.c Mon Aug 12 17:43:03 2002
***************
*** 1519,1524 ****
--- 1519,1530 ----
ap_log_error(APLOG_MARK, APLOG_DEBUG | APLOG_NOERRNO, r->server, "Expiry date calculated %ld", (long)expc);
}
+ /* Accommodate BUSTED IIS servers. Microsoft can't seem to pay attention to RFC 2616.
+ * When sending a response of code 304 (HTTP_NOT_MODIFIED) you cannot send any
+ * entity-header variables. This patch deletes only the Content-Length variable. */
+ if (r->status == HTTP_NOT_MODIFIED) {
+ ap_table_unset(resp_hdrs,"Content-Length");
+ }
/* get the content-length header */
clen = ap_table_get(resp_hdrs, "Content-Length");
if (clen == NULL)
