Re: [css-masking] clip-rule

On Mon, Dec 16, 2013 at 3:30 PM, fantasai <fantasai.lists@inkedblade.net>wrote:

> On 12/13/2013 11:27 AM, Dirk Schulze wrote:
>
>>
>> On Dec 13, 2013, at 8:09 PM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
>>
>>  On Wed, Dec 11, 2013 at 7:54 AM, Dirk Schulze <dschulze@adobe.com>
>>> wrote:
>>>
>>>> Hi,
>>>>
>>>> On Dec 11, 2013, at 11:46 AM, fantasai <fantasai.lists@inkedblade.net>
>>>> wrote:
>>>>
>>>>    7. clip-rule is not defined. At all. There's some syntax, an
>>>>> explanation
>>>>>      of how the two values differ, and an example, but no definition
>>>>>      of what the property actually *does*.
>>>>>
>>>>
>>>> I put the explanation into the example [1]. Examples are normative.
>>>>
>>>
>>> Examples are informative.
>>>
>>
>> Thanks for the clarification.
>>
>> The definition of clip-rule is still complete. fill-rule is normatively
>> referenced. It is said that UAs must follow what fill-rule does and there
>> is a normative paragraph that clip-rule just applies to descendants of
>> the <clipPath> element.
>>
>
> I can infer that 'evenodd' and 'nonzero' on 'clip-rule' are interpreted
> analogously to 'fill-rule', but nothing says how it is interpreted for
> clipping. Nothing says that it applies to the clipping path, for example,
> as opposed to some other polygon.
>
> You need a real definition here.


Reading the 'clip-rule' definition [1], it's unclear why it applies to HTML
and most SVG elements since it should only affect SVG path operators that
are uses as a clip.

Maybe this should change to say:

applies to: SVG shapes <http://www.w3.org/TR/SVG/intro.html#TermShape> and text
content elements<http://www.w3.org/TR/SVG/intro.html#TermTextContentElement>
that
are descendants of a <clip-path> element


The text following the property definition seems to confirm this.

1: http://dev.w3.org/fxtf/masking/#the-clip-rule