This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The validator fails on <http://giftfile.org/>, saying it doesn't know how to process an xml+rss file. This site has both an index.html file and an index.rss file, but to my knowledge content negotiation is disabled for the .rss file (due to IE issues). In any case, the validator should probably be putting some weight on text/html and application/xhtml+xml MIME types via the Accept header, as if it were an HTML browser.
% netc giftfile.org 80 HEAD / HTTP/1.0 Host: giftfile.org HTTP/1.1 200 OK Date: Wed, 02 Jun 2004 18:25:25 GMT Server: Apache/2.0 Content-Location: index.rss Vary: negotiate,accept TCN: choice Last-Modified: Sat, 29 May 2004 00:21:37 GMT ETag: "5c138-cb9-35084640" Accept-Ranges: bytes Content-Length: 3257 Connection: close Content-Type: application/rss+xml Or see http://cgi.w3.org/cgi-bin/headers?url=http%3A%2F%2Fgiftfile.org *** This bug has been marked as a duplicate of 18 ***
> *** This bug has been marked as a duplicate of 18 *** I saw that bug, but don't agree it is a duplicate. I don't want to select a custom Accept header, I want validator to do the right thing by default, which is to request a resource as if it were an HTML browser, and put proper weight on HTML types. By the way, I have disabled content negotiation for the directory index at the example URL given, so it will no longer exhibit the problem.
I believe "doing the right thing by default" will be implemented as a side effect of resolving bug 18, as the validator does not send any Accept header nowadays. Perhaps this and bug 785 are not strictly duplicates of bug 18, but very much related anyway; maybe a blocker should be set instead.