Bug 50434 – Marquee elements do not stop animating when scrollAmount is set to 0.

• Status: RESOLVED
• Resolution: FIXED
• Component: Layout and Rendering

[Andy Estes]

Created attachment 75455 [details]
Test case

http://trac.webkit.org/changeset/73189 exposed a DOM interface for marquee elements. A marquee element should not animate when its scrollAmount content attribute is set to 0, but setting scrollAmount to 0 in script does not cause the element to stop animating. Instead, it behaves as if scrollAmount was set to 1. See the attached test case.

[Andy Estes]

<rdar://problem/7863203>

[Andy Estes]

Created attachment 75460 [details]
Patch

[David Levin]

Comment on attachment 75460 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=75460&action=review

A few comments.

> WebCore/rendering/RenderMarquee.cpp:297
> +        int increment = abs(m_layer->renderer()->style()->marqueeIncrement().calcValue(clientSize));

Why not just change the 1 to a 0?

> LayoutTests/fast/html/marquee-scrollamount.html:13
> +        // Marquee's default scrollDelay is 85ms, so wait for 100ms to ensure

Why no change the scrollDelay to 0 and then wait a shorter time?

100ms doesn't seem long but it adds up over lots of tests.

[Andy Estes]

(In reply to comment #3)
> (From update of attachment 75460 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=75460&action=review

Thanks for the feedback David.
> 
> A few comments.
> 
> > WebCore/rendering/RenderMarquee.cpp:297
> > +        int increment = abs(m_layer->renderer()->style()->marqueeIncrement().calcValue(clientSize));
> 
> Why not just change the 1 to a 0?

It seems more concise to remove the call to max() entirely since an absolute value can't possibly be less than 0.

> 
> > LayoutTests/fast/html/marquee-scrollamount.html:13
> > +        // Marquee's default scrollDelay is 85ms, so wait for 100ms to ensure
> 
> Why no change the scrollDelay to 0 and then wait a shorter time?
> 
> 100ms doesn't seem long but it adds up over lots of tests.

Good idea. I can mke that change before landing.

[Andy Estes]

Committed r73398: <http://trac.webkit.org/changeset/73398>