Anne van Kesteren

Resolving fragment identifiers or not?

Sjoerd Visscher explains the problem in still bugs in the implementation of HTML hyperlinks?! I guess that everyone working on a project with links at the moment should take a look (Atom working group anyone?) and after that make sure the specification requires following RFC 2396. And for anyone with enough spare time and knowledge about how to hack Mozilla: Bugzilla Bug 275689 - same-document references should not be resolved at all.

I guess Safari and Opera should get a similar bug report. If nobody beats me to it, I will take a look at those browsers as well when I got time.


  1. Konqueror (KHTML engine, same as Safari) does it correct.

    Posted by praseodymium at

  2. Safari 1.2 doesn't.

    Posted by Sjoerd Visscher at

  3. I checked out the latest Atom syntax spec and it points to RFC 2396 bis. I totally missed that one!

    Posted by Sjoerd Visscher at

  4. This is exactely the problem I encoutered when I first implemented the [target=] tag on Gathering of Tweakers since we also use a base HREF for images. I had to work around that using javascript :/

    Posted by crisp at

  5. Is there, with the current (incorrect) implementation a possibility to POST a form the same URI, without using the absolute URI, or the URI relative to the base URI? form action="" is relative to the base URI, I guess. action="#" looks like the only solution to me... Are there other ways?

    Well, if there are no other ways, I find it hardly imaginable that the bug mentioned will be closed with a WONTFIX.

    Posted by Robbert Broersma at