My original page just used the version of mootools hosted on Google
AJAX API.

Here is the same page (on 1.3 compat) but on JSFiddle.  Broken in
IE6-7 for me.

http://jsfiddle.net/7PymN/

(I didn't have any luck with no-compat mode either).

Phil

On Oct 25, 4:07 pm, johnwait <[email protected]> wrote:
> Here's what I get in the debugger for IE 7.0.5730.13:
>
> http://oi54.tinypic.com/2yzar6q.jpg
>
> Hope this helps. I don't currently have access to any other Windows
> systems to test this on.
>
> On Oct 25, 6:56 pm, johnwait <[email protected]> wrote:
>
>
>
>
>
>
>
> > That's weird.
>
> > Slick actually finds the element with id 'sauce' using IE's native
> > getElementById(), but then rejects it because the id specified
> > ('sauce') doesn't match the 'id' attribute of the found element on the
> > document's DOM tree (which still gives out 'cloneMe').
>
> > That looks like an IE bug to me, as even after changing the id of the
> > element (which itself is confirmed to have worked because
> > getElementById() returns the element by the new id), but the DOM still
> > contains the old id for the element.
>
> > Here's the break down of what happens:
> >         // string '#sauce' is stripped of the hash, so that value for id is
> > 'sauce'
> >         // this.getElementById() finds and returns in el the cloned element
> > with id 'sauce'
> >         var id = expression.substring(1), el = this.getElementById(id);
> >         // checks if we need to bail out because nothing was found, but it's
> > not the case here
> >         if (!el) return found;
> >         // special check for IE, because IE might return element by name,
> > instead of by id.
> >         // so we check against the DOM tree to see if the id matches the
> > corresponding attribute
> >         if (local.idGetsName && el.getAttributeNode('id').nodeValue != id)
> > return false;
> >         // and false is returned, because 
> > el.getAttributeNode('id').nodeValue
> > returns 'cloneMe'
>
> > So, basically, el.getAttribute('id') gives out 'sauce' but
> > el.getAttributeNode('id').nodeValue still gives 'cloneMe'.
>
> > Jonathan R.-Brochu
>
> > On Oct 25, 2:57 pm, Phil Freo <[email protected]> wrote:
>
> > > I've had about half of IE7 users tell me it works, and half say it's
> > > broken.  It might even have to do with a minor release of IE or the
> > > version of JS inside IE.
>
> > > What's the full IE7 version you're testing on?
>
> > > I've been able to reproduce the bug so far on:
> > > IE 7.0.5730.13IC
> > > IE 7.0.5730.13.xpsp.080413-2111
>
> > > On Oct 25, 11:35 am, Oskar Krawczyk <[email protected]> wrote:
>
> > > > Tested on IE7/xp - works fine (the alert appears).
>
> > > > O.
>
> > > > On 2010-10-25, at 20:13, Phil Freo wrote:
>
> > > > > Hello,
>
> > > > > I've found a pretty bad bug in Mootools 1.3 that didn't exist in 1.2, 
> > > > > that affects IE6-7 users.
>
> > > > > I've mentioned this on IRC and not everybody is able to reproduce it, 
> > > > > but I have had several people be able to reproduce it on different 
> > > > > computers running IE6-7, and it has been the root cause of a lot of 
> > > > > user complaints recently on Quizlet.com.
>
> > > > > The follow pages "work" if you get an alert('sauce') popup.  They 
> > > > > don't work if you get a JS error.  (NOTE: Only IE6-7 are affected).
>
> > > > > Simplified snippet of the bug:
> > > > > On 1.2 (this work's fine, alert() msg) ->http://quizlet.com/moo2.htm
> > > > > On 1.3 (this is broken in IE6-7, js error, no alert()) 
> > > > > ->http://quizlet.com/moo3.htm
>
> > > > > Screenshots from browsershots.org:
> > > > > On 1.3 (IE6, broken) 
> > > > > ->http://browsershots.org/screenshots/8dac1f27fb13eb66ae567b4cbe86bb1f
> > > > > On 1.3 (IE7, broken) 
> > > > > ->http://browsershots.org/screenshots/5750950ec55a32673d4595bdabc9ffbf
> > > > > On 1.2 (IE6, works) 
> > > > > ->http://browsershots.org/screenshots/ca52437ac6722955a23b7d729f994ce7
> > > > > On 1.2 (IE7, works) 
> > > > > ->http://browsershots.org/screenshots/f9870b41a948a5c6521289e40e047c13
>
> > > > > There's definitely a problem... (I've reproduced it on several 
> > > > > different real PCs with native IE7 installations, as well as a couple 
> > > > > different virtual machines with IE6).
>
> > > > > What's the next step here?
>
> > > > > Thanks,
> > > > > Phil

Reply via email to