I don't think it's a Webkit specific issue. It's a common issue.

 

What I suggested was like below:

* The below 2 is equivalent to..

body { background: red }
p { background: url("chess.png") 40% / 10em gray
       round fixed border-box; }


* By current spec (http://dev.w3.org/csswg/css3-background/#the-background) :
body {
    background-color: red;
    background-position: 0% 0%;
    background-size: auto auto;
    background-repeat: repeat repeat;
    background-clip: border-box;
    background-origin: padding-box;
    background-attachment: scroll;
    background-image: none }
p {
    background-color: gray;
    background-position: 40% 50%;
    background-size: 10em 10em;
    background-repeat: round round;
    background-clip: border-box;
    background-origin: border-box;
    background-attachment: fixed;
    background-image: url(chess.png)


* By my suggestion :
body {
    background-color: red;
    background-position: 0% 0%;
    background-repeat: repeat repeat;
    background-clip: border-box;
    background-origin: padding-box;
    background-attachment: scroll;
    background-image: none }
p {
    background-color: gray;
    background-position: 40% 50%;
    background-size: 10em 10em;
    background-repeat: round round;
    background-clip: border-box;
    background-origin: border-box;
    background-attachment: fixed;
    background-image: url(chess.png) }

 

My 1st example page (http://book.coforward.com/sample/css3/12_css3_background-size.html)
 was a page to explain the "background-size" attribute to people.


But the author uses "background" after using "background-size"
 because he didn't know "background-size" is a part of background shorthand.


He may test his page before the "background-size" comes into the "background shorthand",
and after the browser updated, he'll think "The browser have a new bug".

 

I know my suggestion looks weird compared with other background attributes,
 and I agree with the contents modification seems better.


But I don't know how to make all the contents providers modify their sites
 for our customers who think our browser have a bug.

 

------- Original Message -------

Sender : fantasai<fantasai.lists@inkedblade.net>

Date : 2012-10-04 22:35 (GMT+09:00)

Title : Re: [css3][background] background-size as the background shorthand

 

On 10/03/2012 10:03 PM, Kyung-Tae Kim wrote:
> Hi all,
>
> What I want to suggest is how about not to initialize background-size
> when the background shorthand doesn't include it.
>
> The 'background-size' attribute was added to the background shorthand by CSS3 spec.
> (http://dev.w3.org/csswg/css3-background/#the-background)
>
> Then, it was applied to the Webkit several months ago.
> (https://bugs.webkit.org/show_bug.cgi?id=27577)
>
> However, many legacy sites set background-size first and then set background for setting other background attributes.
> In these cases, the 'background-size' is initiaized, and the backgrounds become weird.

Sounds like a WebKit-specific legacy problem. I don't think we should
make everyone else change just because WebKit had a bug. If the sites
in question were coded to the standard and tested across platforms,
this problem would not be showing up.

In general, WebKit should make sure that when it introduces a new
property that is defined to be set by a shorthand, even if that
property's values are not yet parsed as valid values of the shorthand,
it should still be reset by the shorthand. This is easy to forget,
so WebKit reviewers should be careful to check for that when approving
patches for new properties. (Same goes for other implementers, too,
of course.)

~fantasai