Re: Mutation events - slowness examples

I've made some changes to make this demo more realistic:
1. use document.body.clientHeight to force a rendering update.
2. insert HTML with more complex layout. The original HTML 
"<span>A</span>" repeated 1024 times had a extreme ratio of 
DOMNodeInserted events to layout complexity.
3. make the insertion point 5 DIVs deeper in the document.

My results (approximately):
Firefox: 40ms -> 50ms
Webkit: 28ms -> 31ms
Opera: 35ms -> 45ms

This is still a shallow and wide HTML fragment being inserted - 128 
DOMNodeInserted events for not much content.
In the real-world I would imagine that the non-JS overhead of 
DOMNodeInserted events is almost always irrelevant to the UX.
The JS cost will be incurred no matter what form of handling is provided.

The test-page is attached.
I'll endeavour to provide realistic tests for other Mutation Events in 
the next few days.
Sean


François REMY wrote:
> Result on my computer :
>
> Test 1 (something is done by JScript)
>    IE:                     Not Applicable
>    FireFox:           30 > 160 (x5 or +130)
>    Chrome:           5 > 60 (x10 or +55)
>    Opera:              15 > 45 (x3 or +30)
>
> Test 2 (nothing is done by JScript, just for comparaison)
>    IE:                     Not Applicable
>    FireFox:           30 > 80 (x3 or +50)
>    Chrome:           5 > 40 (x8 or +35)
>    Opera:              15 > 30 (x2 or +15)
>
> There's cleary two factors that apply :
> ==> Time needed to move from managed C++ to unmanaged JScript
> ==> Time of JScript execution
>
> These two factors are clearly dependant of the browser. In FireFox, it 
> seems there's clearly a problem with that.
> Others browsers react smartly enough.
>
> Regards, Fremy
>
> --------------------------------------------------
> From: "Michael A. Puls II" <shadow2531@gmail.com>
> Sent: Saturday, June 06, 2009 3:49 PM
> To: <www-dom@w3.org>
> Subject: Mutation events - slowness examples
>
>> Are there any example pages of mutation events being really slow and  
>> causing things to lag?
>>
>> Are they always slow in every browser?
>>
>> -- 
>> Michael
>>
>