Re: some case-insensitivity cases to test from Boris Zbarsky on 2012-12-12 (www-style@w3.org from December 2012)

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Wed, 12 Dec 2012 12:37:51 -0500
To: www-style@w3.org
Message-ID: <50C8C0EF.7050100@mit.edu>

On 12/12/12 9:10 AM, Leif Halvard Silli wrote:
> I checked the above page. And mark element test, where the k has been
> replaced with a  kelvin-K, is buggy. The test says "highlighted if
> matched". And in Safari, it gets highlighted. However, when you look at
> the DOM in the Inspector, then it is evident that it is not the kelvin
> K but the normal, lower-case k that is used in the element. (So
> javascript apparently replaces the kelvin K with the normal k.)

What presumably happens here is that createElement lowercases its input 
per spec, and this lowercasing in WebKit is not ASCII-only (per spec it 
should be!), leading to the observed behavior.  So that's still a 
somewhat useful test, but testing the selector in WebKit probably 
involves an explicit createElementNS with the non-ASCII name, since that 
won't do case-folding during element creation.

-Boris

Received on Wednesday, 12 December 2012 17:38:21 UTC