Re: [csswg-drafts] [css-display] Define interaction of display:contents and replaced elements

My comment on the call was that we had chosen ```display: contents``` 
to behave like ```display: inline``` instead of ```display: none``` 
because that was how Firefox layouted the controls (it also seemed to 
have the preference of most authors in the room). 

I didn't check what the getComputedStyle function returned, but to be 
fair -- in Edge at least -- this function could return you any value 
independently of what we actually store in our formats, I don't think 
that is particularly relevant. If Gecko desires to store formats in 
one way that is not what a spec says the computed value is, it should 
be perfectly fine as long as there is some code in getComputedStyle 
that takes care of returning the expected value to authors, and they 
have function in their codebase like 
```shouldComputedXyzabcBeAuto(style, boxSource)``` to adhere to 
existing spec text that conditions on if the computed value of Xyxabc 
is Auto.

In Edge, getting a value through getComputedStyle will regularly 
trigger some new cascade pass for the asked property if the value 
stored in the formats isn't enough to synthetize the value to return. 
This is desirable also to prevent :visited styles from leaking out for
 properties that might be affected by that.

In short, I am not entirely sure why there is so much push from gecko 
to "dumbify" the computed value of properties to the point they end up
 being the specified value in almost all the cases. The computed value
 is a concept, and is not tied to your implementation of formatting in
 any strong way.

-- 
GitHub Notification of comment by FremyCompany
Please view or discuss this issue at 
https://github.com/w3c/csswg-drafts/issues/540#issuecomment-282062764 
using your GitHub account

Received on Thursday, 23 February 2017 17:33:12 UTC