W3C home > Mailing lists > Public > public-fx@w3.org > October to December 2012

Re: Specifying mask type for <mask> elements

From: Brian Birtles <bbirtles@mozilla.com>
Date: Wed, 03 Oct 2012 09:05:34 +0900
Message-ID: <506B814E.2000304@mozilla.com>
To: public-fx@w3.org
(2012/09/27 11:41), Dirk Schulze wrote:
>> I'm ok with adding the keywords back, but I don't think we should ignore
>> them.
> Ok, so we add 'auto' to the mask-type property, but keep 'luminance' as default?

Dirk and I continued some of this discussion on IRC to work out what the 
sticking points are.

Assuming that we allow keywords like "alpha" etc. when a mask property 
points to an SVG <mask> element, one situation where we had different 
expectations was the following:

   <mask mask-type="alpha" id="a">
     ...
   </mask>
   <g mask="url(#a) luminance">
     ...
   </g>

My expectation here is that when 'a' is applied to the <g>, its 
luminance values are used.

This expectation is based on the pattern that the specific overrides the 
generic. In this case the mask property is more specific than the 
properties defined on the <mask> element since it corresponds to a 
specific application of the <mask>. The same would be true even if the 
mask property appeared in a style block. It's still a specific 
application of a re-usable mask.

If I understand correctly, others suggested that the 'luminance' in this 
case should simply be ignored?


Note that we're not talking about the default value of mask-type here. 
That's luminance. We're also not talking about the behaviour of "mask: 
url(#a) auto" or "mask: url(#a)". That's also agreed upon. I'm just 
asking about the behaviour of the mask property when it specifies 
luminance or alpha AND points to an SVG <mask> element. Another example 
being:

   <mask id="a">
     ...
   </mask>
   <g mask="url(#a) alpha">
     ...
   </g>


The other issue of how to define this so we can support SVG stacks 
(which currently does not seem possible with the way CSS Images is 
defined and complicates the processing here) is something I've yet to 
look into.

Best regards,

Brian
Received on Wednesday, 3 October 2012 00:06:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 3 October 2012 00:06:04 GMT