Re: cdata, javascript and xhtml1.1

David,

glad to see you're one of us, this thread was at html, however ig also 
has a place, though you may also want to post your response to html...
Its fairly inaccessible if you need a second degree to write it, well 
as I say I'm having problems :-]
I'm having real problems putting up an xhtml page that a browser will 
treat as xhtml.

I'd just as soon have a simple example, with a .htaccess file, 
javascript, css, and an xhtml file.

http://www.peepo.com/alfi-x/music-x/w-x.xhtml
is my current best effort, which validates at w3.  IE6 and Mozilla both 
want to download this rather than open it.

.htaccess file: AddType text/xhtml+xml xhtml

<meta http-equiv="Content-Type" content="text/xhtml; charset=us-ascii" 
/>
your point about the meta tag is well taken, thanks, a number of other 
respondents missed this:

document.write('<bgsound src="#" id="mysound" loop="1" />')
the IE specific javascript bgsound not being quoted, the code is mainly 
cross-browser,
however as mentioned elsewhere, I've had problems for some years 
getting sound-on-event to work without flash in mozilla, icab, opera or 
other browsers. this need is not yet recognised by w3

these files are transient, as I change them in response to replies, and 
my understanding :-]

thanks again

Jonathan


On Sunday, November 24, 2002, at 06:35 PM, David Woolley wrote:

>
>> Could someone expand on why CDATA is essential?
>>
>> our page* validates fine without it, yet when I include it the page no
>> longer appears to load :-( ie5.2 mac mozilla20021115 mac ie6 pc
>
> The script is within an XHTML comment.  The CDATA markup prevents the
> recognition of this as a comment.  The HTML DTD declares script content
> to be CDATA, but XML requires that a parse be possible without the
> DTD, so CDATA must be indicated explicitly in line, and cannot be in
> the DTD.
>
> Any browser that honours the script is not treating the resource as 
> XHTML,
> which is probably reasonable as neither the HTTP header nor the meta
> element claim that it is.
>
> Note that it should not validate, as it contains elements that are not 
> in
> the DTD, and has missing quotes which make it not even well formed.
>
> There's a broken fragment link (src=#).  There are bogus javascript:'s 
> at
> the start of the event handlers - I guess this parses as an unused 
> label.
>
> Is some browser really treating meta reply-to as a sneaky link element,
> and accepting URLs.  Even if it were a valid http-equiv, the standard
> RFC use of Reply-To headers does not include a mailto: scheme prefix!
> There is a perfectly good link element notation for this, although it
> dropped out of the specs because no major browser implemented it 
> (Mozilla
> now does):  <link rev="made" href="mailto:xzxx@kfadf.example.com">.
>
> The scripting appears to be IE specific.
>
> Not sure what this has to do with accessibility.
>

Received on Monday, 25 November 2002 04:23:30 UTC