| Every now and then I try to fix stuff, or at least identify what needs fixed... |
| September 20, 2006 | Computers and Software |
| A while back I was bit by wierd self-closing tag attempts in HTML. I ran into an HTML page that used <a id="myid" />, which makes sense for anchors as "targets I link to within a document", but not so much for anchors as 'enclosing content that links to some other document or target". Not that I see any way to change that overloaded tag type at this stage. Recently, the fine folks at Surfin' Safari wrote about the oddity that is "valid XHTML 1.0" as interpreted by web browsers, and how it's oddly broken because it usually just gets read as "nearly valid HTML 4.0.1" So http://www.whiterose.org/test/testanchor2.html is a page that validates as XHTML 1.0 Transitional and serves as text/html. The anchor tag with the misguided self-closing "/" confuses the hell out of all the browsers I saw, and is similarly broken on all browsers I tried. To be honest, I'm not even sure if it should just self-close or not. The wierdest thing is that Safari and Konquerer display different behavior than Mozilla family/IE6 if the self-closing anchor tag is inside 2 nested divs, but not just 1. I ran into this on Making Light and tried to report it as a Safari bug (8879), but it was closed as badly reported. I'd be happy to have help describing it better, because while it would be nice if it worked right, it should at least not work too differently from Mozilla and IE. This still happens differently with Safari revision 16490 and Firefox 2.0b2. Anyone want to help me write a better bug report? | |
| .:Posted by Michael on September 20, 2006 11:36 PM:. | |
One issue with your bug report is that your reduction is very large, and doesn't make it particularly obvious what the problem is. The HTML 4 specification states that the closing </a> tag is required, so if you omit it the browser is left to guess at where the <a> tag should be closed. It would be preferable if Safari and Mozilla handled this the same inside multiple levels of DIV-nesting. I would suggest that you refile your bug report, and address only a single issue in it. It would be very helpful if your test case was reduced to the minimum possible code that demonstrates a difference between Safari and Firefox. Having the test case and bug report describe concisely and accurately what you expect to see versus what you actually see would also be great.
Thanks for taking the time to come up with this test case and bug report! If you need any further assistance with this, please feel free to drop in to the WebKit IRC channel -- there are always plenty of helpful people around.
| .:Posted by Mark Rowe ( total) on September 21, 2006 12:04 AM:. |
| Archives |
| Archives | ||||
| Category Archives | ||||
| Recent Entries | ||||
| Comment Leaders | ||||
| Links | ||||
|
|
||||
 |
||||
| Syndicate this site (XML) | ||||
|
||||
