Bug 222516 – Blending of border-image-width should be discrete between "auto" values and other types

• Component: Animations
• Status: RESOLVED FIXED

[Antoine Quint]

Blending of border-image-width should be discrete between "auto" values and other types

[Antoine Quint]

Created attachment 421752 [details]
Patch

[Antoine Quint]

Created attachment 421753 [details]
Patch

[Dean Jackson]

Comment on attachment 421753 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=421753&action=review

> Source/WebCore/ChangeLog:9
> +        got tricked by an "auto" Length returning true for Length::isZero(). So we also check that 0 values aren't

Is this a bug? Seems like returning isZero for an auto length is wrong. Did you check why that is?

    if (isCalculated())
        return false;
    return m_isFloat ? !m_floatValue : !m_intValue;

Seems like auto should be similar to calculated. But I bet some code would break if we made a change here.

[Simon Fraser (smfr)]

Comment on attachment 421753 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=421753&action=review

> Source/WebCore/animation/CSSPropertyAnimation.cpp:756
> +            if (a.type() == b.type() || (a.isZero() && !a.isAuto()) || (b.isZero() && !b.isAuto()))

Wouldn't it be better to add a function to Length that does what isZero() && !isAuto() does?

[Antoine Quint]

(In reply to Simon Fraser (smfr) from comment #4)
> Comment on attachment 421753 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=421753&action=review
> 
> > Source/WebCore/animation/CSSPropertyAnimation.cpp:756
> > +            if (a.type() == b.type() || (a.isZero() && !a.isAuto()) || (b.isZero() && !b.isAuto()))
> 
> Wouldn't it be better to add a function to Length that does what isZero() &&
> !isAuto() does?

Actually, I think isZero() should return false when isAuto() is true, as Dean suggested. I'll file a patch to see if EWS is happy with that.

[Antoine Quint]

Created attachment 421771 [details]
Patch

[Antoine Quint]

(In reply to Antoine Quint from comment #5)
> (In reply to Simon Fraser (smfr) from comment #4)
> > Comment on attachment 421753 [details]
> > Patch
> > 
> > View in context:
> > https://bugs.webkit.org/attachment.cgi?id=421753&action=review
> > 
> > > Source/WebCore/animation/CSSPropertyAnimation.cpp:756
> > > +            if (a.type() == b.type() || (a.isZero() && !a.isAuto()) || (b.isZero() && !b.isAuto()))
> > 
> > Wouldn't it be better to add a function to Length that does what isZero() &&
> > !isAuto() does?
> 
> Actually, I think isZero() should return false when isAuto() is true, as
> Dean suggested. I'll file a patch to see if EWS is happy with that.

This looks like a winner, there are a bunch of new PASS results from css-grid tests as well!

[Antoine Quint]

Created attachment 421773 [details]
Patch

[EWS]

Committed r273635: <https://commits.webkit.org/r273635>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 421773 [details].

[Radar WebKit Bug Importer]

<rdar://problem/74835871>