Bug 199410 – background-clip:var(--a) invalidates -webkit-background-clip:text when --a:text

• Component: CSS
• Status: RESOLVED FIXED

[Jim Ratliff]

See http://dabblet.com/gist/062ef5b3cef1557ed994aae73ee6ff8a for reduced test case.

This problem occurs when `background-clip` is set to `text` using a CSS variable. It causes WebKit to ignore a valid `-webkit-background-clip:text` and instead reverts to `background-clip:border-box`.

Note that the following works fine (renders background through clipped text):
  -webkit-background-clip: text;
  background-clip: text;

But the following does not work (computes `background-clip:border-box` rather than `background-clip:text`) in WebKit, Safari, and Chrome (but works fine in Firefox):
  --a: text;  */
  -webkit-background-clip: text;
  background-clip: var(--a);

Apparently, WebKit finds `background-clip: var(--a);` invalid at computed-value time (even though it is equivalent to `background-clip:text`, which works fine) and causes WebKit to throw out the preceding `-webkit-background-clip: text;` statement.

[Alexey Proskuryakov]

*** Bug 199409 has been marked as a duplicate of this bug. ***

[Radar WebKit Bug Importer]

<rdar://problem/52595653>

[Tyler Wilcock]

Created attachment 410721 [details]
Patch

[Tyler Wilcock]

With this patch, all permutations of:

 * Properties `-webkit-background-clip` and `background-clip`
 * Values `-webkit-text` and `text`
 * Values as CSS variables

clip correctly.

[Tyler Wilcock]

Created attachment 410735 [details]
Patch

[Darin Adler]

Comment on attachment 410735 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=410735&action=review

> Source/WebCore/css/parser/CSSPropertyParser.cpp:3164
> +    if (RefPtr<CSSPrimitiveValue> value = consumeBackgroundBox(range))

auto

[Darin Adler]

Comment on attachment 410735 [details]
Patch

Ideally we would also add a test about the integration with vat that was originally reported.

[Darin Adler]

(In reply to Darin Adler from comment #7)
> Ideally we would also add a test about the integration with vat that was
> originally reported.

Never mind. That doesn’t make sense. The original report was just confused about cause I think.

[Tyler Wilcock]

Created attachment 410753 [details]
Patch

[Tyler Wilcock]

(In reply to Darin Adler from comment #6)
> Comment on attachment 410735 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=410735&action=review
> 
> > Source/WebCore/css/parser/CSSPropertyParser.cpp:3164
> > +    if (RefPtr<CSSPrimitiveValue> value = consumeBackgroundBox(range))
> 
> auto

Someday I'll get this right :)

Thanks Darin.  Updated the patch — I'll mark it cq? once the tests are green again.

[EWS]

Committed r268158: <https://trac.webkit.org/changeset/268158>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 410753 [details].