[whatwg] usemap="" and related issues

[Ian Hickson]

On Wed, 26 Nov 2008, Jonas Sicking wrote:
> Ian Hickson wrote:
> > On Thu, 26 Jun 2008, Jonas Sicking wrote:
> > > > On Sat, 18 Aug 2007, Jonas Sicking wrote:
> > > > > Since ID is case sensitive everywhere else, I don't see a reason to
> > > > > make
> > > > > an exception from that rule here. That seems to unnecessarily
> > > > > complicate
> > > > > implementation as well as introduce weird inconsistencies for authors.
> > > > It already is inconsistent for usemap="". At least for legacy Web
> > > > content I don't think we can do much about it. At that point, I'd rather
> > > > just extend that to XHTML than to keep another difference.
> > > In mozilla for HTML we only look at the name attribute, and only do so
> > > case insensitively. For XHTML we only look at the id attribute, and are
> > > always case sensitive.
> > > 
> > > We have had a number of bugs filed on id not working on HTML, (with most
> > > of them pointing at the XHTML spec as a reason it should work) but they
> > > all use the same casing for the usemap attribute and the id attribute.
> > > 
> > > Do you have any data showing that using case sensitive matching for the id
> > > attribute would break compatibility with any pages?
> > 
> > I do not. It seems like something where being incompatible with what IE does
> > is unnecessary, though.
> 
> I just did a little bit of testing, but it seems like IE *always* treat id's
> in a case insensitive manner, including for getElementById. If we are
> duplicating that quirk then we should do it consistently, not just for image
> maps.
> 
> However I don't think we should.

Ok, changed it so id="" only works case-sensitively. name="" is still 
compatibility-caseless.


> > > What I did notice in our code though is how we deal with the case when
> > > there are multiple <map>s with the same name. In this case we generally
> > > use the first <map>. But if the first <map> is empty, we use the first
> > > non-empty <map>. This was done for compatibility with some sites. See
> > > https://bugzilla.mozilla.org/show_bug.cgi?id=264624
> > > 
> > > I have no idea if this matters today or not.
> > 
> > I couldn't reproduce this behavior.
> 
> Try the following markup in firefox:
> 
> <map name="foo"></map>
> <map name="foo">
>   <area shape=circle coords="10,10,10" href="http://www.mozilla.com">
> </map>
> <img src="http://www.mozilla.org/images/feature-logos1.png"
>      usemap="#foo" width="20" height="20">
> 
> You'll note that the second image map is used. However if you insert any 
> <area>s in the first image map, then the first image map will be used.
> 
> I don't know if this is important or not these days, I'd love to remove 
> this behavior, but at least when the above bug was filed there were 
> sites depending on it. Data would be great.

On Fri, 28 Nov 2008, Lachlan Hunt wrote:
> 
> This only seems to occur in quirks mode, not in standards mode.  But 
> neither Opera, Safari or IE8 have the same behaviour.  Additionally, the 
> site reported in the bug you mentioned no longer suffers from the bug. 
> Therefore, it doesn't appear to be necessary that we should require that 
> behaviour.

Based on Lachlan's comment, and your desire to remove this quirk, I 
haven't added this to the spec.

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