Massa, Harald wrote:
> I applied the patch to mod_scgi.c, recompiled, installed...

> and it works. Great!!!

That's great news! It's nice to see those 304's in your log. ;-)

>
> BUT: the IMS-Patch was not enough for mod_scgi.c. Additionally it was
> necessary to add a line with "last modified" header ... and that was the
> main problem generator (the missing one)
>
> I sent the patch to mod_scgi.c to Neil, also include it in this mail for the
> list, if somebody stumbles upon the same error.
>
> *** mod_scgi.c  Sat May 31 21:19:42 2003
> --- mod_scgi.c.patched  Sat May 31 21:17:51 2003
> ***************
> *** 219,224 ****
> --- 219,226 ----
>         add_header(t, "CONTENT_TYPE", lookup_header(r, "Content-type"));
>         add_header(t, "DOCUMENT_ROOT", ap_document_root(r));
>         add_header(t, "HTTP_ACCEPT", lookup_header(r, "Accept"));
> +         add_header(t, "HTTP_IF_MODIFIED_SINCE", lookup_header(r,
> "If-Modified-Since"));
> +         add_header(t, "HTTP_LAST_MODIFIED", lookup_header(r,
> "Last-Modified"));
>         add_header(t, "HTTP_REFERER", lookup_header(r, "Referer"));
>
>         hdrs_arr = ap_table_elts(t);


Interesting... "Last-Modified" is a response header, it really shouldn't appear
in a request...

I took a look at the error log entries you sent me the other day, here's an
example:

 > [Fri May 30 19:51:13 2003] [error] [client 172.16.20.42] scgi: error reading
response headers

This would suggest that mod_scgi didn't like our code setting the Last-modified
_response_ header. I don't understand how the patch above could fix that, since
it deals with _request_ headers. But there could be some aspect of the mod_scgi
code at work that I don't understand.

Congratulations on getting this to to work, Harald!

-- Graham