[whatwg] [Selectors4] case-insensitive attribute value matching (in XML)

Ian Hickson ian at hixie.ch
Tue Jun 5 17:04:55 PDT 2012


On Sat, 21 Jan 2012, Leif Halvard Silli wrote:
> Ian Hickson on Fri Jan 20 14:31:01 PST 2012:
> > On Tue, 26 Jul 2011, Christoph Päper wrote:
> >> Anne van Kesteren:
> >> > I'm still trying to get HTML and browsers to change so that attribute 
> >> > values always match case-sensitively, irrespective of markup language. 
> >> > The current magic attribute list in HTML whose values needs to be 
> >> > matched ASCII case-insensitively is just ugly.
> 
> > The spec changed recently in response to Anne's efforts here. If this is 
> > an area of interest, I encourage you to study the specification to see if 
> > the current requirements are satisfactory.
> 
> The matching rule for attribute names and element names, [1] doesn't 
> match reality, see demo: [2]
> 
> * Gecko uses ASCII case-insensitive matching (as specced by HTML5)
> * Trident/Webkit/Presto use Unicode caseless matching (variant).
>   (Legacy Firefox 3.6 behave like Trident/Webkit/Presto too.)
> 
> The differences affect @data-* and @x-* (and other extensions). 
> Shouldn't spec match Trident/WEbkit/Presto?
> 
> [1] http://dev.w3.org/html5/spec/links#case-sensitivity
> [2] http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1307

On Fri, 3 Feb 2012, Anne van Kesteren wrote:
> 
> The HTML parser only lowercases A-Z so that behavior is somewhat 
> surprising. Quick testing shows it also happens in the DOM (in 
> Presto/WebKit at least). I think it should be treated as a bug in 
> Trident/WebKit/Presto given how the HTML parser behaves, personally.

I agree with Anne. Unless there's a compelling compatibility reason, the 
ASCII variant seems better. It has the advantage of being more performant, too.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'


More information about the whatwg mailing list