[ic] /special_pages/missing.html

[Carl Bailey]

On May 20, 2007, at 12:01 PM, Frank Reitzenstein wrote:

> Subject: [ic] /special_pages/missing.html
> To: interchange-users at icdevgroup.org
>
> Hello,
>
> I am running Interchange 5.2 on Fedora Core 5
>
> I have searched for posts regarding missing pages because I am coming 
> to
> the conclusion if the server doesn't return 404 for a page which is no
> longer on the site, then the search engines go on indexing many more
> pages than are visible on the site, for a very long time.
>
> We are told that Google Page Rank gets distributed amongst our pages,
> and hence in this case to pages which don't exist. The danger is with 
> PR
> being so much diluted, that the vast majority of pages end up in
> "Supplemental Results".
>
> My page /special_pages/missing/html now consists of only:
>
> [tag op=header]Status: 404 missing[/tag]
>
> and I am pleased that finally the search engines will record that the
> missing pages no longer exist. It will also help me when I use the new
> Google webmaster tool to remove pages. However at the moment the user
> sees a blank page instead of a standard 404 message.
> /
> /I notice that a few people have discussed this, and some have found a
> solution. I don't use /Mod::Interchange/ that I am aware. From memory
> there used to be only a private version for Debian users.
>
> I could spend hours trying to hack something. I am hoping that someone
> has an easy answer ;)
>
> Regards,
>
> Frank Reitzenstein


We tried this out, and found that we could get the "normal" missing 
page to display if we also added:

[tag op=header]Content-type: text/html[/tag]

Then apache would serve the page with the correct status code and the 
page displayed exactly as expected.  Before adding the content-type, 
some browsers wanted to download the file returned by the server and 
save it to disk because they received a content type of 
application/x-executable.  The only thing unexpected was that the 404 
does not show up in the apache error log file, only in the access log.  
This makes sense insofar as Apache successfully finds and returns a 
page.  The 404 is not truly an apache error, hence no entry in the log. 
  This might confuse some log analyzer statistics packages.

We are using IC 5.4.1 with Apache 1.3.33 and no mod_interchange.

Regards,
Carl
+ - - - - - - - - - - - - - +
| Carl Bailey
| Triangle Research, Inc.
+ - - - - - - - - - - - - - +