Closed Bug 468768 Opened 13 years ago Closed 8 years ago

Ads that used to load properly, causes webpage flash to refresh itself when ad loads


(Core :: Layout, defect)

Windows XP
Not set





(Reporter: visualbrains, Unassigned)




(Keywords: regression, testcase, Whiteboard: "regression" from 390425)


(1 file)

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv: Gecko/2008102920 Firefox/3.0.4
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv: Gecko/2008102920 Firefox/3.0.4

My blog didn't have any issue till I updated to 3.0.4 
My Ads are intentionally delayed and called in later using changenode()
This worked fine previously and is still working fine in IE 6.

However once I changed to 3.0.4, when the ads load in, it will cause the webpage to blink and reloads all flash files on my web.

Please take a look at my webpage

Reproducible: Always

Steps to Reproduce:
1.go to my webpage
2.wait for ad to load
3.notice it blinks when it loads
Actual Results:  
Webpage reloads each time ad loads

Expected Results:  
It shouldn't have reloaded my webpage
Yeah, this looks really bad. This is a regression from Firefox 2.
Regression range: 2007081516 - 2007081516 (both the same ID).
Bonsai link:
So possibly caused by Bug 390425.
When it first regressed the color of the flashes was different, first green, later on the flashes became deeper.
Blocks: 390425
Component: General → Layout
Keywords: regression
Product: Firefox → Core
QA Contact: general → layout
Version: unspecified → Trunk
The bug occurs both on Windows XP and Vista.
Flags: blocking1.9.2?
Flags: blocking1.9.1?
Attached file testcase
Thanks for the regression range, Ria.
Ever confirmed: true
Keywords: testcase
So... this used to "work" due to a bug, as far as I can tell: removing the <div> didn't reframe the containing block.  You can see the bug by comparing the rendering of 

data:text/html,<span id="a" style="border: 1px solid green">a<div></div>b</span><script>document.body.offsetWidth; var n = document.getElementById("a"); n.removeChild(n.childNodes[1]);</script>

in Gecko 1.9 (correct rendering) and Gecko 1.8 (incorrect rendering).

So in 1.9 and later, we reframe the containing block on this change, and in the testcase the containing block is the body.  Since plug-ins are loaded from frames, you get the observed behavior.

Marking dependent on the "load plug-ins from content" bug, but as a workaround you could put a <div> around the <span> in question.  Or in general not put blocks inside inlines (esp. since that's not valid HTML).

For what it's worth, I don't think this should block.
No longer blocks: 390425
Depends on: 90268
Whiteboard: "regression" from 390425
Flags: wanted1.9.1-
Flags: blocking1.9.1?
Flags: blocking1.9.1-
Flags: wanted1.9.2-
Flags: blocking1.9.2?
Flags: blocking1.9.2-
Is this still an issue? Plugins have long since been moved to content and a few bugs have been fixed to ensure reparenting of plugins is handled according to the spec.
Flags: needinfo?(visualbrains)
Whiteboard: "regression" from 390425 → [closeme 2013-09-08] "regression" from 390425
Resolved per whiteboard
Closed: 8 years ago
Flags: needinfo?(visualbrains)
Resolution: --- → INCOMPLETE
Whiteboard: [closeme 2013-09-08] "regression" from 390425 → "regression" from 390425
You need to log in before you can comment on or make changes to this bug.