Blending of border-image-width should be discrete between "auto" values and other types
Created attachment 421752 [details] Patch
Created attachment 421753 [details] Patch
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.
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?
(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.
Created attachment 421771 [details] Patch
(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!
Created attachment 421773 [details] Patch
Committed r273635: <https://commits.webkit.org/r273635> All reviewed patches have been landed. Closing bug and clearing flags on attachment 421773 [details].
<rdar://problem/74835871>