Bug 18240 – Safari's scrollbar disappears and frame's scrollbar stops functioning on resize because of position:absolute

• Component: CSS
• Status: RESOLVED WORKSFORME

[jasneet]

I Steps:
1. Open Safari.Do not maximize the window, maybe less than half of screen in size.
2. Go to http://bet.hkjc.com/
3. Click on the bottom tab that is by itself with green background. (see 1st step : attached screenshot)
4. Click on the 6th link (see 2nd step : attached screenshot)
5. Resize browser, making it smaller than previous setting.
6. Navigate to the right of the page by holding down the browser's horizontal scroll bar and dragging to the right (see 3rd step : attached screenshot)

II Issue:
Safari: the frame loads with vertical scroll bar on right. If you navigate to the right (holding down the horizontal scrollbar and dragging to the right), you can see that the browser's scrollbar disappears and only the frame's scrollbar appears- for which you cannot click to navigate to the bottom part of the text (you may need to navigate right to left a couple of times by using the horizontal scroll bar to observe that the vertical scrollbar does not respond) Occasionally, there appears to be paint issues that shows up too.

In IE: only the browser's scrollbars appear. Frame does not have a scrollbar.

III Conclusion:
The position:absolute property of the 2nd <iframe> is causing the issue.
FF: working ok *
*working in FF. The vertical scrollbar on the 1st frame (that FF has and not IE) is actually the expected behavior since scrolling is not specified.

To Reproduce:
Open testcase in Safari. Maximize window.
Decrease Safari window's size so the horizontal scrollbar shows on the bottom.
Scroll to the right using the bottom horizontal scrollbar.

IV Other browsers:
IE7: ok
FF3: ok
Opera9.24: ok

V Nightly tested: 31386

[jasneet]

Created attachment 20239 [details]
screenshot

[jasneet]

Created attachment 20240 [details]
reduction

[Robert Hogan]

I can't recreate this. Is it still a problem?

[Tab Atkins]

The referenced website no longer appears to demonstrate the problem, and I don't see any problem with the testcase.

I'm marking this closed.  Please reopen with a new testcase that shows the problem more clearly if it's still a problem.