Re: UA's implementation of ::selection

On May 15, 2010, at 2:24 AM, David Hyatt wrote:

> Yeah this seems reasonable to me.  I'll make that change in WebKit.
> 
> dave

Wait; whaaaat?

If I do this...

p { color:#999; }
p::selection { color:black; }

...then I will no longer get the UA text-selection background-color? That does not seem reasonable at all to me.


> 
> On May 15, 2010, at 3:17 AM, François REMY wrote:
> 
>> It seems preferable to not apply a background
>> when no background is specified in the
>> ::selection.
>> 
>> Using the default selection background seems
>> very bad in an accessibility point of view because
>> the developer is unable to determine which
>> color will be used (so, the text may end up to be
>> nearly unreadable in some UA while perdectly
>> readable in some others, due to UA or OS settings.
>> 
>> The solution of IE and Opera is here to use no
>> background for selection when no background
>> is specified. When no text color is specified, the
>> original color of the text is used instead.
>> 
>> 
>> --------------------------------------------------
>> From: "Boris Zbarsky" <bzbarsky@MIT.EDU>
>> Sent: Saturday, May 15, 2010 6:10 AM
>> To: "David Hyatt" <hyatt@apple.com>
>> Cc: "François REMY" <fremycompany_pub@yahoo.fr>; <www-style@w3.org>
>> Subject: Re: UA's implementation of ::selection
>> 
>>> On 5/15/10 12:04 AM, David Hyatt wrote:
>>>> The chosen colors for selection in the absence of any specified pseudo element come from the platform.  We have both active and inactive foreground and background colors (so 4 total possible different colors). If the pseudo element specifies only a background color and not a foreground color, we'll use the platform foreground color (and vice versa).  Any specified selection colors in the pseudo element will be used in both the active and inactive states.
>>> 
>>> That doesn't really explain the behavior I see with "background: none" vs "background: transparent"... (which should give equivalent specified values, note).
>>> 
>>> -Boris 
>> 
>> 
> 
> 

Received on Saturday, 15 May 2010 19:31:10 UTC