CSS2 Test Suite Failures

11th July 2003 · Last updated: 5th October 2016


On Windows XP I ran the latest versions of three major browsers through Eric Meyer's CSS2 Test Suite. That's Internet Explorer 6 (Service Pack 1), Mozilla 1.5 and Opera 7.21. Using the results you can see which CSS commands to be wary of using, until each browser becomes fully compatible. I found all browsers tested suffered from commands that either don't work at all, or work in odd ways. Explorer failed a lot of tests mainly due to its lack of support for inherited values.

I also carefully edited some tests slightly to confirm the results I got were true. Plus I tested Explorer when necessary in XHTML Standards Mode, as this can change its behaviour when using certain commands.

Please note that Netscape 7.1 uses the same code engine as Mozilla 1.4 so the results should be near identical. Also note that the CSS2 Test Suite is still in Draft form - Eric Meyer is adding to it over time. I noted one test had a new addition very late on in my testing, which changed Explorer's result from a pass to a fail! So be aware that the results I have may change if the tests are updated, or even removed. I cannot vouch for my results to be future-proof. I might even make some of my own. (Now how about a CSS3 Test Suite Eric?)

View The Results

(The RED cells indicate a browser has failed half or more of the test. The ORANGE cells indicate a test that's almost correct, but contains one or more remaining faults. The GREEN cells indicate a browser has passed the test.)

Comments (5)

Comments are locked on this topic. Thanks to everyone who posted a comment.

  1. Chris Hester:
    I got it wrong - the negative margin tests were correct - the browsers did not fail (except in one instance). While working on a new layout I was making a line wider than it's container by using negative margins. (So if the container has padding, the line overrides the padding to reach the edges.) Works great! But then I realised that of course negative margins are meant to move a block outwards, not inwards as I'd thought when compiling these results before. So I've changed them so the relevant browsers now pass, not fail.

    Posted on 25th August 2003 at 10:08 pm
  2. Jeff Jackson:
    I wouldn't say that "Nothing happens" with my copy of IE6 SP1 on the 8.3 margin-right test. But it does look like the negative margin doesn't work.

    Posted on 28th October 2003 at 4:24 pm
  3. Anonymous:
    9.9.1 z-index Mozilla Firebird 0.7
    If the body has a z-index >= 0 it seems to work.

    Posted on 13th November 2003 at 5:31 pm
  4. taku:
    p {margin-left: 1.25cm;text-indent: -1.25cm;}

    .hang {float: left; width: 1.25cm;}

    <p><span class="hang"(a) </span>Text Text Text</p>

    Is it bug or normal behaviour of rendering (referring to Mozilla)?

    Would someone be so kind and advise me how to check if this bug is registered already in Mozilla bug tracking system. (or maybe it is already in bugzilla?)

    Posted on 15th September 2004 at 1:05 pm
  5. Ethan:
    Taku, I'm not sure what bug you're referring to. Putting the closing bracket on the "span" tag might help.

    (You have <span ... </span> )
      missing bracket --^

    Posted on 27th September 2004 at 8:06 pm