W3C home > Mailing lists > Public > www-style@w3.org > December 2013

Re: [css-masking] clip-path with invalid url

From: Dirk Schulze <dschulze@adobe.com>
Date: Fri, 13 Dec 2013 19:08:44 +0000
To: Tab Atkins Jr. <jackalmage@gmail.com>
CC: fantasai <fantasai.lists@inkedblade.net>, "www-style@w3.org" <www-style@w3.org>
Message-ID: <298E6D84-EA7F-4B5D-8200-142F9FCCDDD9@adobe.com>

On Dec 13, 2013, at 7:49 PM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:

> On Thu, Dec 12, 2013 at 2:28 AM, Dirk Schulze <dschulze@adobe.com> wrote:
>> On Dec 11, 2013, at 7:00 PM, fantasai <fantasai.lists@inkedblade.net> wrote:
>>> On 12/11/2013 07:46 AM, Dirk Schulze wrote:
>>>> 
>>>> On Dec 11, 2013, at 11:46 AM, fantasai <fantasai.lists@inkedblade.net> wrote:
>>>> 
>>>>>  6. # If the URI reference is not valid [...], no clipping is applied.
>>>>>     Please clarify whether a stacking context is still created
>>>>>     or whether the behavior is equivalent to specifying 'none'.
>>>> 
>>>> I would say a stacking context should be created to match the behavior
>>>> but think that implementations don’t do that currently. I would like
>>>> to base the decision on the current implemented behavior.
>>> 
>>> Then please investigate currently-implemented behavior. But please also
>>> raise this to the WG, as the implementors might decide they don't like
>>> the currently-implemented behavior.
>> 
>> I checked the behavior on Firefox, Safari and Chrome. (IE just supports clip-path and mask on SVG which does not have stacking contexts.)
>> 
>> All implementations create a stacking context for clip-path on HTML even if the url() is “invalid”. Means the fragment identifier does not exists, resource is not loaded or does not point to an <clipPath> element. I also tested the behavior on ‘mask’ with the same result.
>> 
>> In all cases the three engines WebKit, Gecko and Blink do create a stacking context.
>> 
>> Now it is up the implementations if they want to change the behavior. Given that all implementations are consistent, I do not expect that to happen.
> 
> That's the correct behavior anyway - we shouldn't be basing things
> like stacking contexts on used-value time information, which network
> requests qualify as whenever possible.

A question that we had recently on Firefox was: You have a url() that is valid at parse time but actually does not reference anything existent. Is the computed value “none” or the url()? I would again assume the latter but would like to have a clarification.

Greetings,
Dirk

> 
> ~TJ
Received on Friday, 13 December 2013 19:09:15 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:39:17 UTC