Re: [css21] What's the used value of left/right when overconstrained in relpos?



On 4/29/13 4:42 PM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:

>On Mon, Apr 29, 2013 at 4:29 PM, Sylvain Galineau <galineau@adobe.com>
>wrote:
>> For CSS2.1 the change that makes most sense to me is to eliminate the
>>bit
>> about
>> values being ignored and make sure the width prose aligns with the
>>height
>> prose
>> i.e. it consistently states that the specified steps define the used
>>value.
>
>Sounds good.
>
>> Then CSSOM's job is to define the interop behavior of gCS(); here this
>> means
>> defining the set of properties for which it'll only resolve relative
>> lengths
>> and percentages and leave specified absolute lengths alone.
>>
>> Does that make sense?
>
>To be specific, I think we'd define that, if the *computed* value was
>"auto" and the computed value of the opposite property was non-"auto",
>return the negation of the opposite property's used value.  Otherwise,
>return the used value.  Right?
>
>Are we sure the bugwards compat is worth keeping?

I assumed we'd specify the existing interoperable behavior. I think you're
suggesting specify what the Blink patch you mentioned [1] now does, right?
If so, your definition looks right.

I am not sure what the compat impact of making this change would be; it's
not something you can query a large database of raw content for. It's also
something three browsers need to agree on changing at the moment so I would
expect caution/pushback.


[1] https://chromiumcodereview.appspot.com/13871003/

Received on Monday, 29 April 2013 23:52:33 UTC