[Webkit-unassigned] [Bug 39621] Extreme memory growth on DOM Hanoi test

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=39621





--- Comment #6 from Brett Wilson (Google) <brettw at chromium.org>  2010-05-24 17:53:38 PST ---
Chrome this to implement session saving without polling. When the user changes a form field, we set a timer (to avoid saving too often) and then serialize the current form state to disk. Then we can restore the form state when recovering from a crash.

The other form elements do the same thing, but they derive from HTMLInputElement which is where the code lives for those controls.

It's not immediately clear why this call has to be in the constructor. My guess is that I was thinking if a form control was added with some initial value, we would want to count that as a change to the form state of the page. However, it appears that the initial set is coming from setNonDirtyValue. So I think the call to notifyFormStateChanged in the constructor can be safely removed.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.