437383 - onmouseout does not fire when navigating back to page

Status: NEW

(Core :: DOM: UI Events & Focus Handling, defect)

Description

[Steve Chapel]

Attachment: Canyonero's testcase

Reproducible: Always

Steps to reproduce:
1. Open testcase
2. Click on Google link
3. Click the Back button

Expected results: After navigating back to testcase, link should appear green

Actual results: After navigating back to testcase, link still appears red

This was mentioned in the MozillaZine forums by Cyberkilla and Canyonero wrote the testcase. <http://forums.mozillazine.org/viewtopic.php?t=659342>

Comment 1

[Jordan]

Anybody picked up on this yet?

Comment 2

[Olli Pettay [:smaug][[e-mail]]]

Is this a regression? If it is, regression range would be great.

Comment 3

[Jordan]

(In reply to comment #2)
> Is this a regression? If it is, regression range would be great.
> 
I'm assuming that means finding the build where this started happening.

I am afraid I have absolutely no idea, Smaug.
It has been showing this behaviour for a long time. Months and months.

I didn't report it because I did not have a Bugzilla account at the time, and I thought that somebody else would have noticed it.

It is a regression, however - it worked in Firefox 2.

Comment 4

[Jordan]

Can somebody confirm this for me? It's pretty easy to confirm(thanks Steve).

Comment 5

[Patrick S]

Confirmed with Canyonero's testcase on:

Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.1a1pre) Gecko/2008061012 Minefield/3.1a1pre

Comment 6

[Olli Pettay [:smaug][[e-mail]]]

If someone had time to find the regression range ...
Nightly builds are here http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/

Comment 7

[Steve Chapel]

Firefox 2.0.0.13 exhibits the same behavior as Firefox 3, so this is not a recent regression. Exactly which version of Firefox 2 worked?

Comment 8

[Jordan]

I left Firefox 2 at 2.0.0.12.

I'll check for you, just the make sure.

Comment 9

[Jordan]

Okay, I have tested Firefox 2.0.0.12, and the problem is not there.
Let me elaborate on that...

If you :
- click the link
- move mouse away
- return to previous page(without moving mouse)

The box stays red. If you move the mouse even a single pixel, the onmouseout event is triggered.

In Firefox 3, you can move the mouse anywhere(except over the box again) and it will never fire the onmouseout event.

This is a regression because it means that javascript tooltips, 'mouseover' image links(and generally anything that changes the state of the page prior to you visiting another one) will remain when they should not.

Tooltips follow the mouse out of the boundaries of the element, whereas they should have been hidden again by the onmouseout event.

Even Firefox 2 is not perfect - it probably should check for mouse events the moment you return to the page - but I never noticed the issue until the Firefox 3 behavior.


It is clearly related to the way they freeze the state of the page before changing to another one.

Comment 10

[Jordan]

I'm not sure if there is anything more I can do about this.

I will stop bumping it and wait until there is more of a response.
You no doubt have other priorities at this time.

I personally see this as a very irritating regression which affects a lot of sites(and not just a minority of strange circumstances), but it's clearly too late to discuss a fix before Firefox 3 is released now.

Comment 11

[Jordan]

Any update on this? I haven't a clue how to make this more prominent.

It's a lot more important than some of the other bugs that seem to get a lot more attention around here.

Comment 12

[Steve Chapel]

It's hard to believe that it's an important bug to fix when there's only one vote and no one has given any actual website that has a significant problem due to this bug. Please do not comment again unless you have useful information to provide <https://bugzilla.mozilla.org/page.cgi?id=etiquette.html>.

Comment 13

[Olli Pettay [:smaug][[e-mail]]]

If someone could try to find the regression range, that'd be great and would
significantly increase the probability that this gets fixed soon.