[whatwg] CSS canvas() function

[Charles Pritchard]

On 12/2/2010 2:48 AM, whatwg-request at lists.whatwg.org wrote:
> Message: 2
> Date: Wed, 01 Dec 2010 21:58:46 -0500
> From: Boris Zbarsky<bzbarsky at MIT.EDU>
> To:whatwg at lists.whatwg.org
> Subject: Re: [whatwg] CSS canvas() function
> Message-ID:<4CF70B66.7060009 at mit.edu>
> Content-Type: text/plain; charset=UTF-8; format=flowed
>
> On 12/1/10 6:43 PM, Tab Atkins Jr. wrote:
>> >  My question was specifically for an out-of-document<iframe>.
> Most browsers don't load documents in such<iframe>s at all, right?
Is this one in specs somewhere? It's relevant to some work I'm doing on 
another spec.

We've certainly hit some issues with such thinking in IE, in our 
compatibility work,
in activating an Active X based canvas element. IE does not apply CSS 
browser behaviors until
an element was added into the DOM tree.

As your side has stated, canvas elements outside of DOM require a 
special call.

..

Generally, I agree, <iframe>s should not be used out of document, as we 
have suitable APIs, like XHR.

That said, hackish solutions for corner cases will always exist: [iframe 
style="visibility: hidden; z-index: -1"]
exists to serve. Just as style="opacity: 0" comes in extremely handy for 
some cases. But those users
are free to write that into their scripting / HTML. It's just important 
that we know it's a defined behavior.


>> >  Should this work?  The rendering of a non-seamless<iframe>   doesn't
>> >  depend on any other elements in the document.  In general, any
>> >  replaced element seems to fall into this camp.
> I don't think that's true.  For example,<svg>  is a replaced element in
> CSS terms.  But its rendering depends on stylesheets, media, etc, etc.
Now that SVG is being integrated in with HTML that's certainly true; I'd 
think there was a time
when it wasn't. I agree, "replaced  element" is too broad.