Trying to find the source of an Accept: header problem

Tim,
	The following is everything I know about the Accept: problem. Can  
you show me what's causing my problems?


Yosi

-----

On January 29, 2007, Tim Berners-Lee made the following change to  
cwm's http code. [1] This was labeled with the reason
> Separate Accept: headers for buggy servers
It was noticed yesterday [2] that this fix was broken; you cannot  
have more than one header in python's urllib libraries, and the code  
was changed back.
When asked today about these ``buggy servers'', Tim referenced the  
server configurations given by Best Practices [3], saying they were  
brittle and could not handle Accept: header that allowed for  
something other than rdf/xml, as indeed cwm will accept other things.  
I don't know how to read Apache rewrite rules, so I'm not sure what  
I'm looking for. Perhaps it is the rule [4] saying

# Rewrite rule to serve RDF/XML content if requested
RewriteCond %{HTTP_ACCEPT} application/rdf\+xml
RewriteRule ^example4/ example4-content/2005-10-31.rdf [R=303]

I was told to e-mail public-swbp-wg@w3.org , but I cannot do this  
until I know what the problem is.





[1] http://dev.w3.org/cvsweb/2000/10/swap/webAccess.py.diff? 
r1=1.27&r2=1.28&f=h
===================================================================
RCS file: /sources/public/2000/10/swap/webAccess.py,v
retrieving revision 1.27
retrieving revision 1.28
diff -u -r1.27 -r1.28
--- 2000/10/swap/webAccess.py	2007/01/25 20:26:50	1.27
+++ 2000/10/swap/webAccess.py	2007/01/29 18:54:03	1.28
@@ -5,7 +5,7 @@
  dereferencing a URI to get a document, with content negotiation,
  and deciding on the basis of the Internet Content Type what to do  
with it.

-$Id: webAccess.py,v 1.27 2007/01/25 20:26:50 timbl Exp $
+$Id: webAccess.py,v 1.28 2007/01/29 18:54:03 timbl Exp $


  Web access functionality building on urllib2
@@ -73,7 +73,9 @@
  #        return open_data(addr)
          return urllib.urlopen(addr)   # buggy in 2.4.2 with CStringIO
      z = urllib2.Request(addr)
-    z.add_header('Accept', 'text/rdf+n3, application/rdf+xml')
+#    z.add_header('Accept', 'text/rdf+n3, application/rdf+xml')  # Fine
+    z.add_header('Accept', 'text/rdf+n3')   # Split for some broken  
sites
+    z.add_header('Accept', 'application/rdf+xml')
      if referer: #consistently misspelt
          z.add_header('Referer', referer)
  #    z.add_header('Accept', 'text/plain q=0.1')


=======================================

[2] http://lists.w3.org/Archives/Public/public-cwm-bugs/2007Jul/ 
0000.html
[3] http://www.w3.org/TR/swbp-vocab-pub/
[4] http://www.w3.org/TR/swbp-vocab-pub/#recipe4