[Webkit-unassigned] [Bug 122097] ASSERTION FAILED: !style->propertyIsImportant(propertyID) in WebCore::setTextDecorationProperty

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=122097





--- Comment #14 from Ryosuke Niwa <rniwa at webkit.org>  2013-10-19 07:39:53 PST ---
(In reply to comment #13)
> > We also need tests for that.
> > 
> > > Source/WebCore/editing/EditingStyle.cpp:-1411
> > >          // text-decoration: none is redundant since it does not remove any text decorations.
> > > -        ASSERT(!style->propertyIsImportant(propertyID));
> > 
> > I think the point of this code is that if text-decoration was specified with !important then we don't want to be overriding it with a new text decoration.
> > We should add a bunch of test cases for that.
>   But here we should always override as per current flow.
> Here we are not allowing removing of text property if it is declared !important.this way we would be end applying text decoration in two way simultaneously(one by style property and other by utag)
> 
> Regarding removal of ASSERT in setTextDecorationProperty():
> 
> setTextDecorationProperty() is private function and called from two places in EditingStyle Class. I checked the both scenarios and found ASSERT is wrong in setTextDecorationProperty(). There is no meaning of !important if new text decoration value is empty in Editing scope.

Are you use neither function is called while pushing style in ApplyStyleCommand?  The problem here is that even if new text-decoration was empty or some other value, we shouldn't be overriding a property with !important.

Although we don't do a good job dealing with !important anyway so there is an argument to be made about ignoring !important in some cases.

Regardless, we need a lot more test cases involing !important in applying & removing inline styles.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.