380861 - Page constantly redraws making it impossible to read

Status: RESOLVED DUPLICATE of bug 368603

(Core :: Layout: Tables, defect)

Description

[Steve England [:stevee]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a5pre) Gecko/20070515 Minefield/3.0a5pre ID:2007051504 [cairo]

Go to http://powerball.com/pb_info.asp and observe the crazy redrawing. If you narrow the browser, the page grows in length. If the browser is wide enough, it just constantly redraws downwards before jumping to the top and starting over.

This page works in Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8.1.4pre) Gecko/2007050204 BonEcho/2.0.0.4pre

Comment 1

[Steve England [:stevee]]

Works:
- Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/2006120606 Minefield/3.0a1

Weird Behaviour:
- Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/2006120804 Minefield/3.0a1

Caused by the reflow branch I expect.

Comment 2

[Eli Friedman]

Hmm, that's funny, seems to work fine in my build, modulo a couple "reflow dirty lines failed" and "invalid divisor" assertions.  (It does eat CPU, but that's because of the stupid JS marquee.)

Comment 3

[-fullmetaljacket-]

wfm too

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a5pre) Gecko/20070516 Minefield/3.0a5pre ID:2007051616 [cairo]

Comment 4

[Jon Baumgartner]

Definitely a problem for me.

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a5pre) Gecko/20070516 Minefield/3.0a5pre ID:2007051608 [cairo]

Comment 5

[Jesse Ruderman]

I was able to reproduce a few days ago (at least at some window sizes), with the content height growing by more than a few pixels per second.  But now the worst effect I see is the content height jumping up and down a few pixels.

Comment 6

[Dão Gottwald [:dao]]

scrollmarquee() does that, per lefttime=setInterval("scrollmarquee()",20). 
With DOMi, you can nicely clear that interval and set a new one with a delay of 1000, for instance.

I don't know why that causes the <td> to grow, though:

if (parseInt(cross_marquee.style.left)>(actualwidth*(-1)+8))
cross_marquee.style.left=parseInt(cross_marquee.style.left)-copyspeed+"px"
else
cross_marquee.style.left=parseInt(marqueewidth)+8+"px"

Comment 7

[Eli Friedman]

Can you retest this with current trunk?  I still can't reproduce this bug.  (And a simplified testcase would be really nice, if you can.)

Comment 8

[Jesse Ruderman]

I still see a "bouncing up and down" effect when I make the window narrow.

Comment 9

[Jesse Ruderman]

Attachment: testcase A

Comment 10

[Jesse Ruderman]

Attachment: testcase B

Comment 11

[Jesse Ruderman]

I lose at Bugzilla.