687054 - StorageEvent is not always fired

Status: RESOLVED WORKSFORME

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

Description

[Alexander Kovalev]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:5.0.1) Gecko/20100101 Firefox/5.0.1
Build ID: 20110707182747

Steps to reproduce:

I have two separate browser windows. Each window listens to `storage` event. 
But `storage` event is not fired always. I expect `storage` event to be fired every time I change the value of `id` property in localStorage

   localStorage.setItem('id', uid);

Each window has its own uid. But the `storage` event is triggred only twice. When the first window updates localStorage, and when the second window updates it for the first time.

Firefox 5.0.1. Mac OS X.

PS
http://jsfiddle.net/alexanderk/sQXsu/show/
This is the demo page which illistrates this bug.
Open two firefox windows with firebug.
Start clicking on "Update local storage" in one window and another.
You will see that `storage` event is fired only twice

I have tested this page in Safari, Chrome, Opera -- everything works fine.




Actual results:

`storage` event was not triggered every time localStorage was modified


Expected results:

`storage` event is supposed to be triggered every time localStorage is modified

Comment 1

[Alexander Kovalev]

I worked around this problem by clearing value of key I'm updating before setting new value.

  localStorage.removeItem('key');
  localStorage.setItem('key', uid);

It looks like previous values are cached

Comment 2

[Vlad [QA]]

Have you tried to see if the issue is reproducing on the latest firefox?
http://www.mozilla.org/en-US/firefox/new/

Comment 3

[Alexander Kovalev]

no, I didn't.

Comment 4

[Vlad [QA]]

Please do so.

Comment 5

[Josh Matthews [:jdm]]

Confirmed in trunk.

Comment 6

[Josh Matthews [:jdm]]

I believe this has been fixed by bug 683316, because I could reproduce this on a build from Sept 22, but not on Sept 25.

Comment 7

[Ian Nartowicz]

Storage event is only fired if the value actually changes.  I don't know if this is correct or not according to the spec.  This is the behaviour I still see.

Comment 8

[Josh Matthews [:jdm]]

I can no longer reproduce the behaviour described in comment 0 (when once I could), so WFM.