80122 - W3C link checker reports links to mozilla.org as broken

Status: VERIFIED FIXED

(mozilla.org Graveyard :: Server Operations, task)

Description

[Jesse Ruderman]

When I use the W3C link checker (http://validator.w3.org/checklink/) on any of 
my pages that link to http://www.mozilla.org/, the link checker reports that 
mozilla.org returned a 404 error.

Here's a page to test the link checker with:
http://www.cs.hmc.edu/~jruderma/mozilla/testlinks.html

Comment 1

[Dawn Endico]

clicking on the above linkx at validator.w3.org causes these entries in
the error log.


[10/May/2001:15:48:55] warning (  258): for host 18.29.1.50 trying to HEAD /,
send-error reports: error opening /e/docs/error.html (No such file or directory)
[10/May/2001:15:48:56] warning (  258): for host 18.29.1.50 trying to HEAD
/credits/, send-error reports: error opening /e/docs/error.html (No such file or
directory)
[10/May/2001:15:49:02] security (  258): for host 194.213.34.219 trying to LINK
/projects/, acl-state reports: access of /e/docs/projects/ denied by ACL (null)
directive 0
[10/May/2001:15:49:05] warning (  258): for host 18.29.1.50 trying to HEAD
/newsbot, send-error reports: error opening /e/docs/error.html (No such file or
directory)

However, those links are valid and work. I believe that validator is just
whining that our server doesn't respond to HEAD the way they want it to.
That's a problem with validator, not us. I'm inclined to invalidate this
but it would be good to either add something to /e/docs/error.html or 
fix the server to use the standard error message.

Comment 2

[Michael Nahrath]

Same behaviour with <http://valet.webthing.com/link.html>

> I believe that validator is just
> whining that our server doesn't respond to HEAD the way they want it to.

The question is: 
Why doesn't the mozilla.org webserver respond to HEAD queries the way other
software expects? 
This is -at least- very uncommon. Is mozilla.org violating the http:-Specs? 
Or are these link-checkers taking incorrect assumptions?

Comment 3

[Terje Bless]

www.mozilla.org is returning a 404 response to HTTP 1.0 HEAD requests. 
The W3C link checker is simply reporting this. So would a browser if it 
tried to do a HEAD on the URI in question -- to check if it had changed, say 
-- and so will anything else that expects a HTTP server to speak HTTP. 

Why in the world doesn't mozilla.org respond properly to HTTP HEAD 
requests?

Comment 4

[Gerald Oskoboiny]

notes from the HTTP/1.1 spec:

    The methods GET and HEAD MUST be supported by all general-purpose           servers.                                                                      -- http://www.w3.org/Protocols/rfc2616/rfc2616-sec5.html#sec5.1.1                                                                                    The HEAD method is identical to GET except that the server MUST NOT         return a message-body in the response. The metainformation contained        in the HTTP headers in response to a HEAD request SHOULD be identical       to the information sent in response to a GET request.                        
     -- http://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html#sec9.4        
(HTTP/1.0 probably says something very similar)

Comment 5

[Gerald Oskoboiny]

er... sorry about the lack of newlines in my previous comment. (I wish
bugzilla had a 'preview' feature) I'll try that again:

notes from the HTTP/1.1 spec:

    The methods GET and HEAD MUST be supported by all general-purpose
    servers. 
      -- http://www.w3.org/Protocols/rfc2616/rfc2616-sec5.html#sec5.1.1

   The HEAD method is identical to GET except that the server MUST NOT
   return a message-body in the response. The metainformation contained
   in the HTTP headers in response to a HEAD request SHOULD be identical
   to the information sent in response to a GET request.
     -- http://www.w3.org/Protocols/rfc2616/rfc2616-sec9.html#sec9.4

Comment 6

[Andreas Franke (gone)]

Making this bug depend on 
Bug 85499: HEAD http://mozilla.org returns "404 Not Found"

Comment 7

[Joe McCabe]

I believe that the HEAD problem can be fixed very easily by simply fixing the
current NES3.6 config (I can provide details if needed). The LINK problem is a
defect in NES3.6 and requires upgrading to NES3.6sp3 (or NES6.1 if I had my way).

Comment 8

[Dawn Endico]

if you can provide details for fixing the current install, that
would be great.

Comment 9

[Jesse Ruderman]

WFM.  The W3C link checker no longer complains about links to mozilla.org.  It
does complain about links to amazon.com, though :)

Comment 10

[Simon Paquet [:sipaq]]

WFM here too.

Comment 11

[Brant Gurganus]

Verified

Comment 12

[Terje Bless]

Resolution should be FIXED; www.mozilla.org is now Apache 1.3.27 instead of the borken NES 
installation it used to be. This was never a issue with the W3C Link Checker; it was always a server 
issue.

(just so anyone searching Bugzilla will get the right info) :-)

Comment 13

[Simon Paquet [:sipaq]]

right, this FIXED

Comment 14

[Simon Paquet [:sipaq]]

re-resolving

Comment 15

[Simon Paquet [:sipaq]]

re-verifying