[Webkit-unassigned] [Bug 64230] REGRESSION (r73385): Marquee with behavior="alternate" is not working

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=64230





--- Comment #22 from Parag Radke <nrqv63 at motorola.com>  2012-01-23 05:29:20 PST ---
> I don't think you need to add a new image for your test. There are numerous images already in LayoutTests that could be reused for this purpose. I also don't think it's okay to add an image of Apple's trademarked logo without permission from the copyright holder.

using a local image for the test case. 

> > Source/WebCore/rendering/RenderBlock.cpp:4936
> > +    if (!isTableCell() && style()->logicalWidth().isFixed() && style()->logicalWidth().value() > 0 && style()->marqueeBehavior() != MALTERNATE)
> 
> Your ChangeLog says that alternating marquees are a special case when calculating logical width, but you don't explain why, and I don't understand myself why this is the case. A previous version of your patch didn't have this check, so I'd like to see more of a "why" explanation.

We need(style()->marqueeBehavior() != MALTERNATE) check as we always need the marquee's actual content width to compute the initial/end position in case of 'ALTERNATE'.
So we need to calculate the logical width in Alternate case even if fixed width is specified as content has to animate between renderBox().right().x() - contentWidth() and renderBox().left().x() + contentWidth().

left                                                     right   
!<-----------[Render box width(may be fixed)]--------------->!
!                                           <-content width->!  
|                                           =================|at t=0
|=================                                           |at t=half period
|                                           =================|at t=full period   

> > LayoutTests/ChangeLog:9
> > +        and smaller testcase execution time.
> You don't need to say "...and smaller testcase execution time." The slower test was never checked in, so nobody will know that some previous iteration of this patch included a slower test case.

OK, done.

> > LayoutTests/fast/html/marquee-alternate-expected.txt:4
> > +PASS on Initial Position
> > +PASS on after half Cycle Completion
> > +PASS on after full Cycle Completion
> 
> The phrases "Initial Position" and "Cycle Completion" do not need to be capitalized.

OK, done.

> > LayoutTests/fast/html/marquee-alternate.html:10
> > +var halfCycle = 75;
> > +var fullCycle = 150;
> 
> The terms "halfCycle" and "fullCycle" are inaccurate. A full cycle of the alternating marquee takes 2ms, not 150ms. 150ms would be 75 cycles. Is it possible for this entire test to run in 2ms instead?

No, it's not possible to complete this entire test in 2ms, i experimented with various combinations but for this latest test case the first move in position of marquee is observed only at 63ms and 2nd at 123ms

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