Re: Mutation events - slowness examples

Boris Zbarsky wrote:
> Sean Hogan wrote:
>> I don't know what would count as a performance sensitive web-app.
>
> Just as a data point, Gecko trunk had a bug earlier today where we 
> accidentally introduce a performance regression on removing table 
> rows.  The bug was easily caught because the regression made removing 
> a table row O(N) in number of rows and it immediately turned out that 
> all sorts of pages out there sort multi-thousand-row tables by 
> reordering the row nodes in the DOM.
Can you provide some URLs?
>
> Which means that they're doing thousands to tens of thousands of DOM 
> mutations.  And the time required needs to stay small (under 200ms, 
> ideally, since that's about whether the user starts perceiving the 
> operation as laggy).
>
But in this case they wouldn't be using Mutation Events.
> That's the sort of thing that we're worried about here, I think...
Sure. Every feature has its costs, and sometimes those costs mean you
don't use it.
I think the page author should make that decision for Mutation Events.

Received on Friday, 26 June 2009 03:42:54 UTC