437386 - setInterval function is still fired 1 time when onunload was already finished in an Iframe

Status: NEW

(Core :: DOM: Core & HTML, defect, P5)

Description

[Sjoerd Mulder]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9) Gecko/2008052906 Firefox/3.0
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9) Gecko/2008052906 Firefox/3.0

Normally the intervals and timeouts are stopped before the unload is finished, this because the onunload might clear / clean variables which are expected to be in the functions of the timeout.

In Firefox 3 this behavior changed, causing error like "bb is not defined" after the page was unloaded, this behavior only seems to happen in an iframe and not when in a normal page.

Reproducible: Always

Steps to Reproduce:
1. See attachted testcase
2. Open the parent.html
3. Press reload
Actual Results:  
It alerts 'How is this possible!?' because of in the function myVar.doIt the myVar variable should always exists, but apparently it's killed in the unload before the timer stops

Expected Results:  
Nothing happens

This is found in the Backbase AJAX Framework.

Comment 1

[Sjoerd Mulder]

Attachment: Testcase

Comment 2

[Bob Clary [:bc] (inactive)]

Attachment: child.html

Comment 3

[Bob Clary [:bc] (inactive)]

Attachment: parent.html (references child.html attachment)

Comment 4

[Bob Clary [:bc] (inactive)]

over to dom0

Comment 5

[Sjoerd Mulder]

Is this going to be picked up before Firefox 3 releases, because i think it's a pretty important bug!

Comment 6

[Blake Kaplan (:mrbkap) (inactive)]

This is interesting. What happens is that we navigate the outer frame to the new document (canceling its timeouts, etc) but don't get around to telling the inner frame that we've navigated until after paint suppression has lifted. So the timeout has that window in which to fire, and does.

One possible way to fix this would be to iterate through sub-frames when we clear the outer window (if we're not saving its state) and suspend timeouts on them until they get cleaned up later, but I'm not convinced that's the best way to go.

Comment 7

[WADA:World Anti-bad-Duping Agency]

Looks to be similar phenomenon to Bug 199430 and Bug 252542(both are Main Window case) after fix of Bug 296639.
> Bug 199430 eventhandlers still fire while a new page is loading
> Bug 252542 The attached events are not removed while reloading the page. JavaScript errors are occuring

Bug 296639 was landed on trunk only, so not applied to Fx 2 but applied to Fx 3.
Bug 199430 : consecutive click of send button, switch to new URL, Main Window 
Bug 252542 : event such as onmouseover, Reload(==same URL), Main Window
This bug   : Timer event by setInterval, Reload(==same URL).
             It sounds to be IFRAME(or FRAME) only issue.

Comment 8

[Firefox Bug Husbandry Bot]

https://bugzilla.mozilla.org/show_bug.cgi?id=1472046

Move all DOM bugs that haven't been updated in more than 3 years and has no one currently assigned to P5.

If you have questions, please contact :mdaly.