- From: Alexander Surkov <surkov.alexander@gmail.com>
- Date: Thu, 22 Mar 2012 22:57:57 +0800
- To: Joseph Scheuhammer <clown@alum.mit.edu>
- Cc: wai-xtech@w3.org, Aaron Leventhal - Code <aaronlevbugs@gmail.com>
> I don't understand that. Does it mean that since FF3.6, FF's accName > algorithm is/was consistent with the ARIA spec, but that this is "luck", due > to a regression? And that regression is going away? That's right. On Thu, Mar 22, 2012 at 11:53 PM, Joseph Scheuhammer <clown@alum.mit.edu> wrote: > On 12-03-22 1:48 AM, Alexander Surkov wrote: >> >> Hi, Joseph. >> ... >> >>> <div id="label">label<span style="display:none" id="innerSpan">this text >>> shouldn't be part of the label</span></div> >>> <input id="input" aria-labelledby='label'/> >> >> no name from innerSpan (since it's not referred by aria-labelledby), >> that goes with spec and Firefox implementation. > > > It has nothing to do with "not referred by aria-labelledby" because, if the > display:none style is removed from innerSpan, then it *is* included in the > accName. Nothing has changed in terms of aria-labelledby. The only change > is whether innerSpan is hidden or not. The two outcomes are consistent with > "skip hidden elements" -- hence, consistent with the spec. And, apparently, > consistent with FF's implementation. > > One more test case since the rule is more fully "skip hidden element *unless > they are referenced by aria-labelledby* ...". Thus, if a span is referenced > by another element, then it isn't skipped even though it is hidden. > > <div id="label3">label <span style="display:none" id="innerSpan2">this text > will part of the label even though it is hidden</span></div> > <input id="input3" aria-labelledby='label3 innerSpan2'/> > > The spec predicts "label this text will part of the label even though it is > hidden". And, FF outputs the same thing. Again, the spec and the > implementation are in agreement. > > The main issue is how a hidden <label> element is handled. As you noted, > the spec says it should be skipped, and my test showed that FF-11 does skip > it. But... > >> You just spot a regression from Firefox 3.6. > > > I don't understand that. Does it mean that since FF3.6, FF's accName > algorithm is/was consistent with the ARIA spec, but that this is "luck", due > to a regression? And that regression is going away? > >> Thank you. >> Alex. > > > Thanks to you as well. > > > -- > ;;;;joseph. > > > 'A: After all, it isn't rocket science.' > 'K: Right. It's merely computer science.' > - J. D. Klaun - >
Received on Thursday, 22 March 2012 14:58:32 UTC