Bug 75319 – getComputedStyle for border is not implemented.

• Component: CSS
• Status: RESOLVED FIXED

[Alexis Menard (darktears)]

getComputedStyle for border is not implemented.

[Alexis Menard (darktears)]

Created attachment 120689 [details]
Patch

[Tony Chang]

Comment on attachment 120689 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=120689&action=review

> Source/WebCore/css/CSSComputedStyleDeclaration.cpp:2146
> +            list->append(getPropertyCSSValue(CSSPropertyBorderTop, DoNotUpdateLayout));
> +            list->append(getPropertyCSSValue(CSSPropertyBorderRight, DoNotUpdateLayout));
> +            list->append(getPropertyCSSValue(CSSPropertyBorderBottom, DoNotUpdateLayout));
> +            list->append(getPropertyCSSValue(CSSPropertyBorderLeft, DoNotUpdateLayout));

Can you use getCSSPropertyValuesForShorthandProperties here?

[Tony Chang]

Comment on attachment 120689 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=120689&action=review

>> Source/WebCore/css/CSSComputedStyleDeclaration.cpp:2146
>> +            list->append(getPropertyCSSValue(CSSPropertyBorderLeft, DoNotUpdateLayout));
> 
> Can you use getCSSPropertyValuesForShorthandProperties here?

Oh, is it because this is comma separated?  Can we add a param to getCSSPropertyValuesForShorthandProperties for this?

[Alexis Menard (darktears)]

Created attachment 120729 [details]
Patch

[Tony Chang]

Comment on attachment 120729 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=120729&action=review

> LayoutTests/fast/css/getComputedStyle/getComputedStyle-border-shorthand-expected.txt:116
> +PASS computedStyle.getPropertyCSSValue('border').cssText is '320px solid rgb(255, 0, 0), 320px solid rgb(255, 0, 0), 320px solid rgb(255, 0, 0), 320px solid rgb(255, 0, 0)'

This doesn't parse when assigned to border. I think it's important that the value returned by getComputedStyle is a valid CSS value.

There are some values where border can't represent the border state (when they are different for each side).  I think we have 2 options:
1) Never return a value for border (what we do now).
2) Check to see if the four borders are the same and return the computed style if they are all the same.  Otherwise return nothing.

[Alexis Menard (darktears)]

(In reply to comment #5)
> (From update of attachment 120729 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=120729&action=review
> 
> > LayoutTests/fast/css/getComputedStyle/getComputedStyle-border-shorthand-expected.txt:116
> > +PASS computedStyle.getPropertyCSSValue('border').cssText is '320px solid rgb(255, 0, 0), 320px solid rgb(255, 0, 0), 320px solid rgb(255, 0, 0), 320px solid rgb(255, 0, 0)'
> 
> This doesn't parse when assigned to border. I think it's important that the value returned by getComputedStyle is a valid CSS value.

This is a good point.

> 
> There are some values where border can't represent the border state (when they are different for each side).  I think we have 2 options:
> 1) Never return a value for border (what we do now).
> 2) Check to see if the four borders are the same and return the computed style if they are all the same.  Otherwise return nothing.

The second option is what Opera is doing.

[Alexis Menard (darktears)]

Created attachment 120870 [details]
Patch

[Alexis Menard (darktears)]

(In reply to comment #7)
> Created an attachment (id=120870) [details]
> Patch

Implemented as discussed in http://lists.w3.org/Archives/Public/www-style/2011Dec/0551.html

It's the option 2 that Tony proposed.

[WebKit Review Bot]

Comment on attachment 120870 [details]
Patch

Clearing flags on attachment: 120870

Committed r103948: <http://trac.webkit.org/changeset/103948>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.