Bug 27417 – Add spread radius support to -webkit-box-shadow

• Status: RESOLVED
• Resolution: FIXED
• Component: CSS

A spread radius component for box-shadow values is described in <http://www.w3.org/TR/2008/WD-css3-background-20080910/#the-box-shadow>.

Created attachment 33030 [details]
Patch

Treatment of corner radii is deliberately different from Firefox’s. <http://www.w3.org/TR/2008/WD-css3-background-20080910/#the-box-shadow> says “ The shadow should not change shape when a spread radius is applied: sharp corners should remain sharp”, and this patch extrapolates this to not increasing the radius. Firefox makes a special case for the 0 radius by preserving it, while growing any positive radius by the spread radius, which does not make sense to me.

<rdar://problem/7072267>

[Anders Carlsson]

Comment on attachment 33030 [details]
Patch

r=me!

Fixed in <http://trac.webkit.org/projects/webkit/changeset/46097>.

[Hajime Morrita]

Hi mitz, 
Trying fix Bug 49913, I find this bug describes the behavior reported 49913 is intentional.
But I have an objection... 

According to the spec [1], "The shadow should not change shape when a spread radius is applied", 
that means, in my understanding, the circle-shaped border makes circle-shaped shadow.
And Mozilla and Opera go that way [2].
I think scaling shadow radii based on spread is the right thing to do because it "optically" correct.

What do you think?

[1] http://www.w3.org/TR/2008/WD-css3-background-20080910/#box-shadow
[2] http://habrahabr.ru/blogs/css/103170/

I think that the language in the editor’s draft has changed in a way that it is no longer possible to reconcile the WebKit implementation with it. I still think the discontinuity at 0 is bizarre.

<http://dev.w3.org/cvsweb/~checkout~/csswg/css3-background/Overview.html?rev=1.237>

[Hajime Morrita]

(In reply to comment #6)
> I think that the language in the editor’s draft has changed in a way that it is no longer possible to reconcile the WebKit implementation with it. I still think the discontinuity at 0 is bizarre.
> 
> <http://dev.w3.org/cvsweb/~checkout~/csswg/css3-background/Overview.html?rev=1.237>

Mitz, thanks for your rapid feedback! I didn't know such a history.
Though I'm not sure whether I understand your concern about discontinuity,
I'll try something at Bug 49726.
(which is the canonical of Bug 49913 I mentioned at previous post.)

By “discontinuity” I mean that the rule applied to a border-radius value of 0 is different from that applied to any positive border-radius value. To me, this makes no sense.

[Hajime Morrita]

(In reply to comment #8)
> By “discontinuity” I mean that the rule applied to a border-radius value of 0 is different from that applied to any positive border-radius value. To me, this makes no sense.
I got it! We cannot increase nor decrease the radii in that case.